Q3 numbers are not what you want to see heading into an IPO.
Again the issue is not is Anthropic an amazing growth story. Zero would debate that.
The issue is this is a probably $200-300B company pretending it’s a $2T company.
Add to that the insane amount of committed spend, a high interest rate environment which means they’ll pay a painful premium to patch cashflow gaps, and a likely “junk” bond rating (per Friday’s news) and that’s why folks are concerned.
It’s also just wrong to keep dismissing the paper losses from convertible notes as somehow not real losses. Once they are a public company this will be actual value taken away from existing public shareholders so they are legit losses not to be ignored. It’s like claiming equity based comp is free money… it’s not, you’re taking it out of the pockets of shareholders.
They may not be $2T company but they are also definitely not a $200-300B one. Their Q1-Q3 revenue is $33B with $17B in Q3. Even if they magically flatline for Q4, that's $50B revenue for the year. There are exceedingly few tech companies with P/S of 6 or less (equivalent to $200-300B valuation with $50B revenue) and we all know Anthropic's revenue will not be anything remotely like flatline for awhile.
> There are exceedingly few tech companies with P/S of 6 or less
I think an open question is if SOTA labs should be priced like software companies have been. Software has been so valuable because the marginal cost approaches zero as distribution increases. Profit/employee at a company like meta prior to massive AI spending was measured in the millions. Meanwhile, the balance sheets of the labs look entirely different - remaining competitive requires continuous and astronomic capital costs, and there's no indication that this goes away. They are perennially being chased by competitors offering the same performance as models mere months old at fractions of the cost.
Their business to me, at least long term, looks a lot more like an airline, railroad, or other extremely capital-heavy industry with very competitive and low margins. None of those trade at 6+ p/s. Particularly on the buyer side, the switching cost to me as a high-AI-spending employee if a cheaper model offers compelling performance is less than deciding that I should ditch my preferred business travel airline.
I think that's why you're seeing so many calls from the leaders of those companies to regulate their competition away.
But like Google and Facebook have a lot of hardware. They aren’t just pure in the cloud (though Google is running other peoples clouds). But yes their hardware demands per dollar of revenue is much lower
Ultimately it's marginal, meaning that every $ of revenue that Google and Meta make have an associated compute cost with it, but scale your revenue down and you scale compute down also. Moreover this compute cost per revenue is fairly low. A small cloud instance equivalent's of compute can generate a ton of business value.
The big foundational labs are not like this. Every model they release takes a ton of data infrastructure and training (pre, train, post) compute. Then inference itself requires compute. Even if the labs sell tokens, each token embodies a lot of training and inference compute. This is a much more capital heavy business, so I can see revenue multiples being smaller than pure SaaS businesses. A model release bust (like Opus 5) ends up leading to big losses because of the necessary training compute spent on the model and the labor spent on training. A bust SaaS product ends up being wound down quickly, and only costs the labor spent on building it and the compute spent running it.
A good analogy is chip companies before all of the AI related compute pressure. These were largely low margin businesses because of the fixed capital cost associated with each new chip gen, architecture, process, etc
Well, it's true that Google has a lab-like structure. Computing the index is a massive continuous "training" job that takes multiple entire clusters (when I worked there at least), and a Google cluster could easily consume an entire rented colo facility. And you need to have crawled large chunks of the web, then done a lot of expensive calculations, before you can serve a single query. Scaling down the crawl/index pipeline is possible to some extent but not much - no different to LLM training.
So I think the reason for the value gap is purely in the proprietary IP. The recipe for making a good LLM is more or less public, unless you buy into the "everyone just distils claude/gpt" angle. The recipe for making a good search engine is private.
The value in Facebook is in the social network that can't be replicated by just buying servers. The value in Google is/was in the secret IP. Hardware costs are low compared to the value of those. Hardly anyone knows how to make a search engine that works as well as Google does, or how to make a social network that draws everyone away from Insta/FB (ok bytedance do...).
Note the huge difference in strategy and outcome!
Google published essentially nothing on how to build a search engine. Huge R&D spend, zero publications, intense culture of secrecy, extremely low attrition to startups or competitors. They did publish some papers on big data systems but only on tech that was many years old - a typical Google "research" paper pre-2015 would announce that the system it described had been running at scale in production for years already. Google also never faced any kind of split or mutiny in which lots of people with the core secrets left the firm to set up a competitor.
With AI however this changed and they published all their research immediately, also their best employees constantly ran away to join startups. The result was predictable: smaller competitors deployed their ideas faster and better than they did, and now Google has no moat anymore.
OpenAI took the classic Google approach and stopped publishing R&D after developing GPT but this didn't help them build margins because:
1. Apparently the core transformer idea is powerful enough that breakthroughs after that didn't matter much. Nobody has been able to develop some killer edge that leaves competitors unable to catch up. Most improvements to transformers have been about cost reduction, not capability improvement.
2. They had a huge split when the Anthropic guys walked out over safety, taking the knowledge of how to build a frontier LLM with them. Result was a peer competitor from day one, whereas Google had at least six years to build edge before anyone realized web search mattered.
Result: Google-style hardware costs but without any secret IP that could generate revenue commensurate with those costs.
> we all know Anthropic's revenue will not be anything remotely like flatline for awhile
The SOTA US labs are one chinese open model away from being abandoned overnight by the masses. Ive argued before, but argue again here, that the current hardware shortages are the main thing keeping this from being a reality currently. If more people were able to run local models for a reasonable all-in cost, we would have a much higher velocity of iteration and improvement in the open source community. But its still too niche while people cant justify the cost, and so they continue to pay the premium for inference.
The limiting factor now is more electricity than hardware. There is hardware sitting idle because data centers can't get grid connections and power delivery contracts. Even for those who willing to run gas turbine generators on site there are severe supply constraints.
OK, but that's irrelevant to Anthropic's IPO. For the next few years at least very few developers are going to do much with local models. The problem that new data center construction projects are running into is electricity availability. In some areas there's no more power available regardless of cost.
You should stop arguing that because it doesn’t make any sense.
If a Chinese company released an open weight model that somehow was so superior to Anthropic and OpenAI then they can just take the model and deploy it, reverse engineer it, use it to improve their models, and/or do any number of things they currently do today which are of increasing value.
Most people don’t want to or have any interest in running or managing models locally.its not just about cost. The market for running models is infinitesimal in comparison to the market for managed products. That’s why everyone uses Office and not Open Office. Nobody is going to run some Chinese open weight model no matter how good it is on their critical banking infrastructure or something.
There will be space for lots of different solutions and many products and types of models including Chinese open-weight models and American ones too once we get around to making money on them.
The game isn’t just “cheapest open-weight model at all costs” which is what you are asserting here.
> If a Chinese company released an open weight model that somehow was so superior to Anthropic and OpenAI then they can just take the model and deploy it, reverse engineer it, ...
> The game isn’t just “cheapest open-weight model at all costs” which is what you are asserting here.
Decide: is the claim that open weight models will be more intelligent than Anthropic/OpenAI SOTA models, or that they will be cheaper?
> Most people don’t want to or have any interest in running or managing models locally
Most people are also not the primary customers of Anthropic! 5 companies make up 25% of Anthropic's revenue, and I bet there a ton of companies spending tens of $M/month on Anthropic/OpenAI that make up much of the rest. These are all medium/big companies that have the means for self-hosting models if they decide that's more convenient. The cost is of course not the whole story here, since Anthropic/OpenAI generally make better models and that allows using them in more situations, where perhaps other models would fail, but this is also a big cost for Anthropic/OpenAI to continue getting ahead, and the moment they stop they lose this advantage. Also in my experience self hosted open weight models are _fast_, which is _very_ nice for all those tasks where you don't need SOTA intelligence to get things done.
Every country in the world outside Europe and United States will adopt/use an open Chinese model because it will cost them relatively nothing, the only thing they have to supply is hardware. And the Chinese certainly can provide that. And so can several other companies outside of China. But those companies won’t be OpenAI, Anthropic or even Nvidia.
Where is their moat? Ultimately, where is their profit? answer? There is none at least there is none that will come close to whatever imaginary valuation Wall Street comes up with, in the best of times, while Wall Street may be able to carry one such company, but two within the same year?
High profit software products have monopoly pricing power from network affects: everyone uses office for compatibility/familiarity with what everyone else uses, everyone uses Facebook because everyone else does, everyone buys advertising buys from Google because that's who everyone sells to because that's who everyone buys from.
No one has shown a network effect for models yet. Models are commodities. Anthropic etc will make the sort of profits one makes selling commodity goods (a couple of percent), plus a perhaps a bit of extra margin from being a few months ahead of open weight models (if that profit can justify the billions in training costs).
I do agree that local models are irrelevant. But local inference and open weights are different things. You can go on openrouter and see that there are dozens of companies selling open weight inference.
The important distinction is SOTA US labs - ANT and OAI - who can not economically survive reverse engineering / fast following PRC models at PRC pricing. Like obviously PRC operators are not going to flourish in US due to geopolitics alone, but that doesn't stop US operators running PRC models, or reverse engineered Murican open models from creaming US SOTA labs on economics eventually. The closer open models are to SOTA in capability the smaller SOTA TAM gets, and that market might not be large enough to fund/maintain SOTA. SWE is one of the highest margin sectors to capture, and open models already shrunk SOTA niche to critical infra and cybersecurity etc then what other high $$$ niche market going to replace? BioPharma maybe? The point is SOTA capitalization depends on large capability gaps that economically justifies 30T TAM - the more domains open models fast follows to compress TAM, less places SOTA has to carve sufficiently profitable moat, and less sustainable maintaining SOTA becomes, especially for dedicated AI labs with no alternative revenue. It's not cheapest open-weight wins, it's cheap and good enough open weight simply murders current SOTA economics because SOTA likely cannot finance at all costs anymore.
The SOTA can't exist in this model because no company or investor will pay for something that will just be undercut.
What you are arguing is that all progress on models (outside of government labs or internal corporate workings) will effectively stop because there won't be funding for it. That seems unlikely to me. You just need to play it out...
US companies/investors go and invest a trillion dollars or something on models. Someone "does stuff" and reverse engineers their model. They go out of business (you said so [1]).
So now what happens?
Well, the US/DARPA/universities/whatever (all countries with interest) will keep working on this stuff and there will be some funding for it but that's about it. Military applications, physics, things like that.
Chinese companies will also fail and it will not be possible to create new SOTA ones because they'll obviously just be reverse engineered too and given away for free or at super low pricing - whatever happens in America must happen in China too.
So who will build new models? There's nobody around to pay for it besides governments or maybe corporate labs. But corporations and governments will protect the models (obviously, as OpenAI and Anthropic do today to varying extents) and keep the weights and stuff closed source. They'll have some productive use for them even if it's not in the download Claude Code and have at it style of B2C/B2B we have today.
Or
Maybe the trillions of dollars of investment going into these companies actually makes sense and the threat of open-weight models isn't what you and others seem to be suggesting it is. Whatever fate we seem to want to assign to America also happens to China here. To the extent they are behind, they're still running the same exact race. If open-weight models undercut companies they'll also undercut Chinese companies.
Just as Open Office exists today, so does Microsoft Office. Microsoft just makes all the money. I'm not sure why some seem to think that they can use the open source comparison here but not also use the open source economics that we've seen play out.
I think as a productive exercise someone should run this argument and explain why Chinese companies won't also be eaten alive by this alleged race to the bottom and have all of their stuff reverse engineered and given away. How can they invest in new AI companies with SOTA aspirations? Whatever argument you want to make for the US applies to China as well, and we can just give away freedom models to all the countries just like supposedly they will.
[1] "The important distinction is SOTA US labs - ANT and OAI - who can not economically survive reverse engineering / fast following PRC models at PRC pricing."
If SOTA finance model is currently unsustainable, then it's entirely economically plausible it will stop. US investors and SV pouring 100s of billions in accumulated $$$ from decade of cheap money printer and tech war chests into one massive hedge is anomalous business model. Did it make sense of TAM is 30T... maybe, but is TAM still 30T? We are seeing currently most profitable SWE niche get creamed, and other sectors likely follow.
>whatever happens in America must happen in China too.
No, because PRC literally could not do what US did... they simply didn't have access to enough compute to dump so much $$$ capitalization for SOTA race. On the other hand, on paper US WITH MORE compute and (currently) CHEAPER compute can put forth cheaper models to undercut PRC open models... but functionally they can't for the same reason cheap PRC operators can't win in US - geopolitics. Undermining PRC AI requires giving PRC cheap US AI on cheap US compute which is sanctioned. But ultimately that is entirely irrelevant to the fact that CHEAPER US models will undercut EXPENSIVE SOTA US models. The undercutting is not symmetrical. US companies can use cheap PRC hammers (or even cheaper US hammers) to break $$$ US SOTA. US who on paper can provide even cheaper hammers to PRC... but cannot sell those hammers... so PRC stuck using cheap hammers... and since PRC never had chance to spend $$$ on SOTA, there's no equivalent valuable things to break even if they had cheaper hammers.
> trillions of dollars of investment going into these companies actually makes sense
We know the context for which those $$$ investment makes sense... if SOTA held meaningful lead to sell high volume of $$$ tokens which was hedge 1-2+ years ago when years of $$$ commitments were made. But doesn't seem to hold anymore, ergo it no longer makes sense. Microsoft had decades to build ecosystem and moat, just like adobe it was, actually hard and high friction to move away. People are switching models all the time. SOTA AI only labs has much less moat than Microsoft or other incumbent tech with decades of overlapping capture mechanisms.
> I think as a productive exercise someone should run this argument and explain why Chinese companies won't also be eaten alive
The productive answer is THEY ARE ALSO BEING eaten alive.
The even more productive answer is PRC commodity AI doesn't have to win for US SOTA AI to lose.
There's just not much meat on PRC AI bones so they are losing billions instead 100s of billions / trillions. Like there is ongoing domestic AI involution irrespective of US SOTA. Literally the primary reason PRC AI was forced to go free was because DeepSeek is billionaire side project who priced his hobby as free which set economics for all other players. Being sanctioned/compute constrained = they figured out how to close SOTA gap on the cheap but also set ceiling on how much they can lose on race to bottom. In so far their SOTA aspirations, again refer to DeepSeek, he takes some couch change from hedge fund gains or investors, and throws it at smart people and hope it's enough. If its not then they'll probably stop too. What PRC not going to do is mobilize resources equivalent to several % of GDP concentrated in few entities for SOTA AI, at least not until their compute is sorted, in which case competitive exposure / risks becomes symmetrical. In the mean time, things are not symmetrical and arguments against economics of US SOTA AI does not apply to PRC commodity AI.
> No, because PRC literally could not do what US did...
I think the fundamental disagreement is you believe that cheaper models can just break the state of the art. This seems false because if it were true then the SOTA can also stop developing and spending on SOTA and just use a cheaper model to build their own better models and leverage their existing compute and energy infrastructure (or even build it out for better and more capabilities later). Anything that is released the SOTA has access to, but it also has access to things that others don't. It's ahead by definition and it's not possible to come out ahead of it because the SOTA will just absorb whatever open-weight/open-source/whatever is released and add to its own lead.
> SOTA AI only labs has much less moat than Microsoft or other incumbent tech with decades of overlapping capture mechanisms.
And those companies are worth a lot of money, continue to execute at high levels, are systematically important, and are by any definition wildly successful. That's my point. You have to apply the open source style argument fairly and equally. If you're going to suggest the cheaper models (which follow this open source software argument) are going to come into being and do all this stuff you also have to acknowledge that they made no money in comparison to the Microsofts of the world and Microsoft won because it made all the money. Whatever open-weight stuff that gets released everybody gets to use it and so tech companies get to use it too and accelerate their advantages. This is what we've seen play out, why would AI be any different when you're using the same economic paradigm as your argument?
> Literally the primary reason PRC AI was forced to go free was because DeepSeek is billionaire side project who priced his hobby as free which set economics for all other players.
American billionaires can also "just set up a side project" and reset whatever economics exist. The crux of your argument is someone can do this and then SOTA isn't economically viable, but you can't escape that this has to be true for everyone, not just the US. If it's true for everyone you also then have to admit the SOTA can't progress because there's no viable economic model - anytime you make an advance in the SOTA someone "just sets up a side project" and/or steals your model or w/e.
The US SOTA losing just means the SOTA has to be equal for everyone or the US is ahead - you can't have it both ways. Progress is practically impossible because you can't make money and you can't obtain a lead so nobody will do any work since their work will just be immediately utilized by someone else.
If what you are proposing here is true, then China is permanently behind. There isn't an economic model to obtain a "lead", yet the American companies get to just take all the work that Chinese companies do (until they go bankrupt not making money) and add it to their lead.
> The productive answer is THEY ARE ALSO BEING eaten alive.
Well you can just run them out of business then. If they spend any money you just spend more until they can't make money and close up shop. Or is someone going to subsidize them indefinitely?
The economics of these arguments that the US SOTA companies are just going to go bankrupt because of cheaper models is something I just don't find compelling. I've been trying. I debate these things so I can learn, but I'm just not seeing a coherent argument on the topic.
I don't think all of these tech companies and investors are stupid. Greedy to a large extent and some will lose their hats? Absolutely. Stupid? I don't think so. Worst case is all the infra is still useful.
> I think the fundamental disagreement is you believe that cheaper models can just break the state of the art.
No their argument is that the existence of a worse quality cheaper product will eat too much of the market for the SOTA company, making the SOTA company less profitable (not necessarily unprofitable) because the number of people who need/want the quality difference supplied by the SOTA product aren't enough to pay to sustain the extra resources required to supply the next iteration of the better product or even if they are then they are not enough to get a respectable ROI on the current investment.
>SOTA can also stop developing and spending on SOTA and just use a cheaper model
Yes, and they will be forced to (if the market shapes out as above) but the point is that then the company won't be getting the returns that justify the valuation being talked about in the IPO and will likely never return the amount of money that has already been invested in it even if it is a profitable company. i.e. If I spend $20,000,000 making a lemonade stand that is a successful best in class stand and it nets me $50 profit a day then it doesn't matter how that as far as lemonade stands go that it's a pretty amazing example. I am effectively never going to see that 20 million again let alone a real profit.
They're not arguing that these SOTA companies aren't better, they're arguing that there won't be a decent return on investment let alone justifying further investment and means everyone will be taking a huge haircut on what is already invested (with the possible exception of early investors who take the bag with the IPO when it inevitably succeeds based on people disagreeing with this evaluation and thinking it a good investment or simply because of index funds being required to buy in)
Cheaper models break financing SOTA pace, PRC budget fast follow currently is catching up. If SOTA slows/stalls slower SOTA pace = instead of PRC 6 months behind, they're 3 months behind, or period of overlapping parity before US SOTA grabs lead again for a few months. There may not be economic model to maintain durable lead, which SOTA economic doom spiral condition. Now if PRC fast follow couldn't close gap, i.e if PRC only fast following due to distillation and US SOTA can technically prevent PRC fast follow, and extend gap then maybe exorbitant SOTA spend sustainable. If not sustainable everyone moves to cheaper cycles. This an issue if SOTA cannot be done on the cheap, which also true for PRC, but PRC not doing expensive SOTA, they're doing cheap fast follow.
>why would AI be any different
It's not, someone going to grab big share of AI... but that someone has to be competitive enough for exquisite lock in, and frankly with enterprise reducing SOTA AI spend due to SOTA AI costs, current SOTA AI ain't it. See last paragraph for hopium.
> someone going to subsidize them indefinitely
Yes. Same way Alphabet/Meta can subsidize AI, China has their Baba and tech giants with alternate revenue streams to subsidize AI, even if not SOTA, and geopolitics will make sure US cannot run PRC AI out of business in PRC and vice versa. But geopolitics not stopping PRC companies running other PRC companies out of business, same with US companies running US companies out of business. In this case host countries tech giants with high margin businesses squeezing sole play AI labs. It's just US SOTA ANT/OAI getting murdered is massive economic contagion risk for US vs Zai/kimi not for PRC.
> find compelling
This really not complicated, last couple years, US SOTA labs signed 100s of billions of compute based on projected demand that they could monopolize enough demand to finance compute capitalization. This not stupid investor prior, because their only competitor, PRC was export controlled, compute constrained and shouldn't close gap. But PRC closed gap, made models good enough and architecture efficient / cheap enough so other players, including those in US, can capture large % of demand WHILE using fraction of compute. SOTA labs now looking at margin compression, losing share to good enough competitors and obligatory (now excessive) capitalization. This bad combo for current US SOTA economics. The numbers suggest this is what is happening. There is chance demand increases past projections and extra SOTA compute will be boon, but again, we are seeing most profitable SWE enterprise curtail not increase demand.
> Worst case is all the infra is still useful
IMO this actually best case for US medium term favour, but also another strike against CURRENT US SOTA. Reality is current SOTA is grossly overpaying for compute / AI premium. Even if OAI or ANT dies, their compute can be picked up for pennies on dollar and as long as PRC behind on compute there is potential for cheaper future US SOTA economics from cheaper hardware since crash will lead hardware price regress towards mean. But that boon is for the next guy.
Here's the US hopium: post current US SOTA crash, there is potentially curve where US has so much cheap discount compute, because current SOTA taking the haircut, and where PRC has not much compute, that US will have sustainable SOTA prospects, i.e. the world where US pays $1 for shovels instead of $10 and can buy $10000 worth of shovels while PRC can only buy $100. That's feasible after current exorbitant SOTA economics bleeds it's patients to death and the next US SOTA host get the transplants and lighter balance sheet, i.e. web 2.0 for SOTA AI. Incidentally more likely "winner" of AI... if I had to guess assuming incumbent FAANG doesn't just snatch up ANT/OAI if they explode - new player/entrant who are near SOTA and eventually SOTA but is priced more competitive than AI 1.0 SOTA, that enterprise will gladly pay for since not at current exorbitant rates that's causing enterprise to bail, but it will be fine because AI 2.0 SOTA can be sustained from 3T TAM instead of 30T TAM. Now PRC might still be able to fast follow, but US SOTA economics completely different if cheap hardware = US premium SOTA cost only 2x more than commodity PRC tokens instead of 20x.
> Most people don’t want to or have any interest in running or managing models locally
When I argue the hardware cost, I should note that I think this is the limiting factor for _improvement_ of these models. But ultimately I dont think most users will self host in that world. They will still pay for inference, but it will be for one of a dozen commoditized vendors who are able to reap the benefits of inference at scale.
It could happen, but it would be an anomaly. In general if a business is to make money it does not let itself become a commodity vendor. You can't go to General Burger and select from one of a Wendy's or Big Mac or Burger King burger. You can't go to Microsoft and buy a Playstation. There are exceptions, of course, but whoever is building this stuff is going to want to make money and becoming a commodity vendor isn't the path toward that goal.
It's the same reason that open source software has been a commercial failure except for a few evangelists or whoever becomes a cool "thought leader" by releasing a new library and then gets invited to conferences to speak. Software was commoditized in many ways, yet all the big software vendors made more money than GOD because they took whatever commodity parts came to exist and monetized them.
I think the models can be improved with either recursive improvement (it would be faster with better hardware but if you can do it at all you just need time) and algorithmic developments. I don't know what those would be, but usually we see improvements to both over time. The algorithms will tend to be trade secrets as well.
Sure, in an ideal Chinese-world, Chinese companies would not be open sourcing their models and weights. But I think its very clearly a deliberate effort to cut off the US-companies at the knees by doing so, as they are unable to compete on the marketplace directly due to the current business/political climate. I also dont think these need to be better than SOTA, just reach a "good enough" bar. Big businesses will continue to use the best-of-the-best, but that will not be the general public and also will not justify current financial outlooks (in my opinion).
Outside of the United States, Europe, Australia, New Zealand, Canada, the rest of the world can choose their computer and choose their model, and they will.
I don't think US companies are being cut off at the knees though. I don't use a Chinese model yet I pay a monthly subscription for Claude. Is there a cheaper Chinese model? Maybe? I don't even care enough to be bothered to look. There are tens of millions of people like me who are in the same boat. Plus you also just have high-end research (as does China) which won't be permitted to leave the country.
The problem here is that let's say it's true that the Chinese are cutting off US companies at the knees. Well the US can just do the same thing back to the Chinese. We might have some money lost and stuff but the data centers will continue to be useful, as will all the infrastructure and energy grid buildout so it's not a total loss (I couldn't care less if accredited investors like myself lose money - you know the risk). Then we just steal Chinese models, do the open-weights thing, and all that stuff. So where does that leave us? Nobody makes any money? (Also where is Europe in all this?)
> I also dont think these need to be better than SOTA, just reach a "good enough" bar.
I agree with you in some respects but I'd like to digress here a bit. Good enough for a general purpose model isn't good enough. You want the best or at least what performs to your expectations. I don't even care about old models nor how much they cost (in my personal life and of which I will be a paying customer as will tens of millions of other people) I just always use the latest and best one unless something else triggers me to use a different one. It reminds me a bit about software updates. Folks around here are like, I'm not upgrading to macOS 17.2 and I'm staying on 17.3! Regular people? Not only can they not tell the difference they don't care - just give me the latest and best thing. period.
On the other hand you can have models that are OK at general purpose tasks but are perhaps specialized for certain applications (software or hardware). Those could wind up being on a spectrum from cheaper (and integrated with lower-quality products kind of like an Android Tv model) to very expensive and high-performant.
I don't see a world in which people are by and large sifting through well today I want to use Kimi 1.1 and well later today I'm going to switch over to Claude Lite 18.7 and dang it's not giving me what I want so I'll switch to whatever model. That's not how people operate - they don't want the cognitive overhead of trying to mess around with something that's good enough. They may have to suffer that based on their economic means, but it isn't going to be an expressed preference outside of very niche categories (like home-brew hackers playing with their toys).
Similarly this is just my opinion as well. I enjoy a good discussion here and I hope you are as well. Positive intents, even if my words my be curt, at times.
> Similarly this is just my opinion as well. I enjoy a good discussion here and I hope you are as well. Positive intents, even if my words my be curt, at times.
I have no issues with this thread : D But back to the discussion...
Again, I elided a lot of thoughts on the whole situation, but the other factor in my belief that this is an inevitability is that the SOTA labs churn too fast in both model behavior but also harness behavior. Normie businesses of the future are not going to hitch their wagon to a moving-target of both behavior and enforcement vs just pinning the equivalent of "opus 4.5" from a run of the mill inference provider. It just doesnt make business sense to be at the whim of a model update. And the entire perpetual motion machine aspect of the SOTA financial outlook _necessitates_ constant churn. If they dont release, they die.
> So where does that leave us? Nobody makes any money?
I think this leaves us with SOTA labs pushing hard for protectionist regulation to essentially require federal blessing to be an inference provider. I also think we will see a _ton_ of dark money poured into research to try and show things like theoretical smuggling/side-channel exfil attacks on models/model weights so that the SOTA companies can paint the _open models and model weights themselves_ as a bogeyman. If they can show that theres potentially latent attack vectors in those paths then that will then prevent all the "middle class" companies from being able to use them, especially if they need to adhere to eg SOC2/HIPAA certification etc
Worldwide, it is will be the cheapest model. That is going to win out. In the United States. and Europe, that will be their (OpenAI/Anthropic’s) markets, every place else probably will be a Chinese Open solution. And that will be no different than EV cars. or solar panels the Chinese dominate worldwide and the same probably will happen with AI models outside the United States, Europe, Australia, New Zealand and one or two other countries.
Also, because of the three hyenas of memory, the Chinese probably eventually will have control of the memory market too in the near 3 to 5 year future.
> Worldwide, it is will be the cheapest model. That is going to win out.
Win on what terms, or in what sense? We've seen this play out via Android versus iOS for example. The cheapest models are more prevalent, but Apple makes all the money. I'd be perfectly happy being the Apple in this case.
It's like you can go and buy a cheap Android phone. It technically does all the same things as the latest iPhone, but it does them worse. So Apple ends up making all the money and delivering the product that people go on to use to do all sorts of other economically advantageous things that make them even more money. Where does that leave the Android phone/model? In the dust, for the most part.
You are misunderstanding a number of things aggressively.
For one thing, one needn't run a model locally, that's a separate--but also very damning--issue.
For another, that's fine if Anthropic and OpenAI can do that, but they will have sunk a bunch of cost into their own training already and people wouldn't pay them what they need when there is always a vastly cheaper substitute that gets the job done.
This is what makes a market. If they keep growing those waiting for the bear case will be left behind. Those who are on the side that the biggest most important AI company will continue to grow revenue will win.
Anthropic has something close to first-mover advantage but that doesn't mean they'll continue to grow revenue. Ten years from now the dominant player could be Google or DeepSeek or some random startup that doesn't even exist yet.
the difference in which side of this bet people take.
of course each stock price is many bets intersecting at the same time (with all their correlations of course)
stocks are claims on future cash flows, currency is claim on future purchasing power, but all can be though of as forms of money, where the prices are based on background information [ie. priors, beliefs, experience] and people's prediction of the future supply and demand.
and the general abstract definition of said bet is asking where are you in the information flow. if you know with sufficient certainty how the people afer you will evaluate some news then you can guess which side to take to make some money.
The only justification for extremely high P/S ratios is if there’s strong evidence that the future margins and profit will be amazing.
With the price of AI rapidly falling to the price of compute, with models therefore basically having zero value, and with Anthropic having rented all its compute from others then there’s little evidence that such high profits are in the future.
There's a world where capabilities narrow and Anthropic is stuck selling a commodity (not like they don't have any competition now). In this world it's not unbelievable that Anthropic would trade around or below 1x revenue no?
I think calling Anthropic a tech company could be like calling an electric utility a tech company.
I think this is focused on the wrong side of the distribution. You can view the value of bear cases at ~zero and it doesn't move the math much.
The $2T comes from the implied odds (perhaps 10 or 20%) that we get near term AGI or RSI and AI becomes the entire economy, crossed with the odds that Anthropic gets there first and is the runaway winner (maybe 25%).
The trouble, in my view, is that today's revenue numbers are essentially irrelevant to the outcome that drives ~all of the expected value. The market has been pretending otherwise, but fundamentally we're looking at a globally top-10 public company that boils down to a binary call option, and the probability underlying that option is inscrutable. That's never happened before, and an inability to divine a fair valuation properly demands extreme volatility, especially in reaction to capabilities progress.
Finally (total speculation): I think Anthropic knows this. I think they're sitting on something with Fable 5.5 and the surprise outperformance of Opus 5.5 is a tell. They intend to release it just before the IPO and make some big claims underwritten by its performance.
> globally top-10 public company that boils down to a binary call option, and the probability underlying that option is inscrutable. That's never happened before
FWIW Bitcoin had the same dynamics. Its valuation between 2011-2021 basically derived from a superposition of two futures: one in which it took over the entire global economy and became the dominant reserve currency, and one in which it failed and its value went to zero. The middle-ground was effectively excluded because currencies have strong network effects and a currency that is not dominant within a market quickly ends up unacceptable to merchants and worthless.
It's still somewhat in that grey zone, used as a form of digital gold by people who believe that the dominant reserve currency (USD) is going to zero as well as a trading medium by rogue nations, but its performance has been lackluster since 2021. It's currently up about 20% since then.
Entirely possible that AI follows a similar trajectory. People assume that it's either going to go bust or take over the economy, but people are really good at overlooking excluded middle grounds. Personally I think it'll settle into a space where it's very effective as a coding assistant, effective at creating propaganda and fake forum posts, and has the effect of killing Google Search (but not really replacing it), but doesn't fulfill all the hype around completely replacing the economy.
> The $2T comes from the implied odds (perhaps 10 or 20%) that we get near term AGI or RSI and AI becomes the entire economy
At this point every churches should IPO and raise hundreds of billions from the implied odds that judgment day is near and Paradise becomes the entire economy, that wouldn't be crazier.
> the probability underlying that option is inscrutable
The "positive" side outcomes are also pretty inscrutable imo: even if AI completely replaced labour, how does it trade against natural resources? Even if anthropic gets there first, might open weights arrive a few months later? If AI drives the value of labour to cents per hour, does that crater global demand?
For sure, the "anthropic stock is a ticket to heaven" scenario is in there somewhere, but even taking complete AGI victory as a given, it seems like a very narrow outcome.
On the third hand: if AGI victory is assumed, even if Anthropic isn't a great bet even under that assumption, where else to put your money? While the music plays, you have to dance.
Another consideration: if we get AGI and it becomes the entire economy, what are the odds that the money you've invested still matters? It seems like both the "universal post-scarcity" or "AI company controls all wealth and power, the peasants live in squalor" scenarios are likely outcomes.
Aren't their operating expenses quite high? That is they depend on high expenses to continuously drive revenue. If they can't keep raising money, then their revenues should be expected go down.
> There are exceedingly few tech companies with P/S of 6 or less (equivalent to $200-300B valuation with $50B revenue)
But a big part of the reason this has been true is that (mostly-software-based) tech companies haven't historically been spending significant billions of dollars every year in capex.
Well, it depends on how you interpret the drop to a 47% quarterly growth rate.
If this is the new normal, that's great news. Revenue growing at 47% per quarter will 10x in 6 quarters. Even if you think it should be a $200B company today, it'll be a $2T company in 18 months, assuming the same revenue multiple.
If you think this is the start of a rapid decline to 0% or negative growth rate, then it's obviously bad news.
The real answer is probably somewhere in the middle -- but exactly where in the middle you think it will fall depends more on your own priors than anything you can read from the data.
> Once they are a public company this will be actual value taken away from existing public shareholders so they are legit losses not to be ignored.
They reported a loss here because they sold notes convertible into shares and their valuation went up. Normally shareholders do not lose money when the stock price goes up.
It's weird that they had to report it like this because they are technically derivatives, and if they had just given out actual shares (as with stock-based comp), they wouldn't. Yes, giving people stock is a real cost, but what happens to the price afterwards is not, and when the price goes up shareholders are definitely making money, not losing it.
Either way, they did report it. It is up to you to decide how you interpret that information.
In some ways it is crazy, but at the same time, anyone that has seen typical efforts of "digital transformation" in aging Fortune 500 companies is going to be skeptical of said company's abilities to become relatively efficient. I always considered that new entrants that don't think like they do would eventually win once the value of putting tech in the middle of decision making, getting rid of large divisions worth of cruft was worth enough. Trying to do the very same thing by competing with everyone is similar once you squint. You get extra money from your parent company to throw around during growth or massive reorg, but still have the problem of iffy incentives we currently see in the modern tech holding company, where what is good for the company, for consumers, and for the random VP have nothing to do with each other.
I don't fully agree with you that it doesn't make sense for Anthropic to claim it is in competition with every other industry.
Take a look at how OpenAI has shaken up higher level math.
Frontier AI labs have a Swiss army knife of a tool that can do as well or better than every other company or collective in every industry.
Say you have a magic hotpot that can deliver any dish you want and each dish it delivers can be as good or better than any restaurant in your area of business. You basically cannot be beaten in the food business. Your competition can only cry about their cooking being more artisanal or whatever but at the end of the day convenience and cost will triumph.
Now say you have a magic pen for art, a magic jukebox, a magic text generator for authorship, and so on. You get the idea why I don't think Anthropic is wrong.
I guess AI will be very competitive in all research fields, but your example is actually interesting: they do not have a magic hotpot, right? So are they competing in the food industry? Or how quickly will they be able to compete with Amazon in logistics?
So I think they can compete in many industries (and I don't like it), but not all of them for sure. They can only compete in all industries with infinite money, and I don't think it would be good for most people.
Anthropic (and OpenAI) has an amazing opportunity to drive the value of multiple other industries to zero, yes. I'm just not quite sure how it benefits them.
Art, music, writing... those are all cheap. They've also already got plenty of good models out there, and people will judge you for using them.
The places where American society needs costs cut or needs an increase in labor - healthcare immediately comes to mind - are places where you'd be asking Anthropic (or any other AI player) to take on massive amounts of liability.
If you screw up someone's drawing, you regenerate the drawing and charge the hack/fraud for the privilege. If you screw up a blood bank order, you pay the survivors a few million dollars. Since you've created a system that can scale blood bank order up-screwing, you'll be doing that a lot. There's not profit in paying out medical torts.
Given that SV-based companies work to dismiss liability even more than most American companies do (which is saying a lot), I don't see that working well for them.
Because if they priced the company more reasonably many later stage investors would stand to lose money. So they’ve backed themselves into a corner of having to go public at a $1T plus valuation just so some later stage investors don’t realize a big loss.
Maybe they can use it, but if there is an overshoot of investment in compute, future competitors may get it for firesale prices while Anthropic pay more than full price. That is, they can get dragged down by those commitments not because their own growth is lacking, but because the market gets weaker. They may even contribute to it, the more competitors they kill, the less demand there is for compute.
To me, "a probably $200-300B company" sounds absurd. You're describing one of the possible outcomes of a massive range that goes from $0, where they get over their skis on financial commitments to well beyond $2T, where their 10x YoY growth continues for a few more years. A $200-300B outcome is possible, but certainly not probable. (And in fact, for us HN commenters, the most probable outcome should be assumed to be roughly what the market will bear.)
>The issue is this is a probably $200-300B company pretending it’s a $2T company.
What do you expect for FY2027 revenue to confidently conclude this? not saying you're wrong, but considering the dramatic changes in their revenue you seem strangely very certain that you know where it will land.
> Okay so Q3 ARR at $67.8bn versus $46.2bn for Q2, +47% QoQ. Turns out tokenmaxxing's end over the summer slowed revenue growth versus an absolutely torrid 100%+ growth rate in Q1/Q2 but 46% sequential is still eye-popping
> Operating margins (GAAP) went from 8% to -3$, with a -28.5% QoQ incremental operating margin. $SPCX compute coming online is the reason.
> Bottom line after seeing this Q3 update: the basic takeaways from the original S-1 don't change, but fears that they were excluding Q3 because it was a disaster were definitely wrong. Do with this information what you will.
It's also worth noting that Q3 only has one week of Opus 5.5 (9/22 - 9/30) which may have changed the economics for Q4.
> fears that they were excluding Q3 because it was a disaster were definitely wrong
That's a very generous reading. Anthropic's entire growth narrative is the rapid growth over the previous 12 months, from near-enough zero to $100bn+ this year. Now we see that actually growth is levelling off and that their revenue was ~$5bn/month over the last quarter, which is barely over their compute commitments, and nothing close to the $100bn+ leaked.
I guess if people thought Q3 would show negative growth rate, sure, it isn't a disaster, but their growth rate slowing so much is surely very bad and matches all the concerns commenters had.
And this is Anthropic! Anthropic are the financial success compared to OpenAI. OpenAI are doing much worse, so this does not bode well for OpenAI either.
People will point out that +46% is still good, but this is like measuring the escape velocity of a projectile that's still at 1000 feet. It's not that the projectile isn't going very fast or won't get very high, it's that if you want to have enough momentum to break out of Earth's gravitation pull you can hardly afford any deceleration.
Really, you're talking about expectations ("expected") of the expected ("price") trajectory (actual business performance).
If the above comment was intending to say "chart go up steeper make price higher, chart go up less steep make price lower", then no dispute there. But that carries virtually zero predictive information about whether the business will achieve "escape velocity" (whatever that means).
You absolutely cannot predict Anthropic's fate based on their growth rate slowing from Q2 to Q3. Obviously.
You can predict the resting place of a projectile with information like this. You cannot do so with a business, especially one as dynamic as Anthropic in an environment as dynamic as today's.
The point being made here I think is that the same can be expressed in natural english language that does not pretend to be a physics class for some reason.
That reason possibly being to grant legitimacy to an otherwise mostly illegitimate and equally mostly fictional branch of the economy (that whole MBA cluster).
Mathematically they can work very similarly though because you have a revenue growth rate, and margin. The question is can that revenue/margin reach break even before you run out of runway (the expected sum of all rounds of investing) .
The main difference I would concede is that gravity, mass, forces, awind drag etc are much easier to model.
Sure they can work very similarly. But they can also not behave very similarly.
Businesses can have stalled growth, degrading margins, etc. etc. and then simply... reverse it. They can raise more money. Their competitors can get regulated out of existence. They can do a literal infinite number of things that cannot be factored into a model like "a projectile."
That's why the projectile analogy is stupid as fuck. It suggests you can presuppose the conclusion by presupposing the conclusion.
It's a model that helps you make present tense valuations of potential futures.
No one said it literally is a projectile. It's an analogy and model. The similarities are in the present tense maths. It's presumed and beyond having to be stated that businesses exist in a complex set of variables all of which could change and might not be as predictable as the valuation maths.
You're not wrong, you're just not adding anything to the conversation that any but the most uneducated do not already know and acknowledge.
And the analogy is literally not useful in any way whatsoever.
It implies that you can know Anthropic's outcome from this slice of data. That is literally not true in any way at all.
Like it's not even remotely or analogously or roughly true.
Anthropic behaves like a high-growth business with a quarter of slowing growth. It does not behave like a baseball. Any conclusions you draw from analyzing it as a baseball will be de facto faulty because the two systems are nothing like one another. The most basic assumption of projectile analysis (that energy is not added in-flight) is the opposite of the most basic assumption of a business (that they are actively profit-seeking).
The analogy adds nothing except ambiguity at best and active falsehoods at worst.
>It implies that you can know Anthropic's outcome from this slice of data.
No it doesn't. All I said was that given how little we have to go on and how early in the process we are, a minor 2nd or 3rd derivative slowdown isn't a positive signal. All indications are they they need additional orders of magnitude in revenue fairly soon.
I also never said that it wasn't self-propelled. I was starting to write about waiting to see if another stage of propulsion comes online but I deleted it.
Okay, well if you weren't suggesting you can predict the outcome based on this slice of data, then "projectile" is a really stupid analogy. The defining characteristic of a projectile is that you can do this.
I think that's an unnecessarily pessimistic take. The way I look at it the industry did not wake up to how big this was until Q4 and so it makes sense that you would see astronomical growth in q1 and Q2 as everybody woke up and started adopting. Seeing the growth Taper down to a 4.5x annualized rate which to your point will only continue to taper but it's still mind-boggling at this scale ... you just don't have anything else like this in the economy. This is like hyperscaler Revenue scale achieved in 15 years being achieved five times as quickly. At this rate Anthropic will exceed AWS Revenue within a year
It is a lot easier to switch model providers than it is to switch cloud providers. And the open weight models from Chinese Labs make Claude a lot more of a commodity than any service provided by AWS.
10x year over year couldn't continue for long--they were what...3.5 years from exceeding US GDP?
47% quarterly growth is still 4.5x annual growth, which would put them at $300+ billion revenue in a year. This is still unprecedented growth for a company of this size.
The negative perspective depends on assuming that the 47% quarterly growth is going to continue dropping soon. And it is going to drop, since it's only a bit more sustainable than last year's growth. The question is how many hundreds of billions they're making a year when it levels off.
> Now we see that actually growth is levelling off ...
I really don't understand how could it be otherwise? They're planning to IP at a valuation representing 5% of the US GDP and, unlike NVidia, they are neither selling shovels in a goldrush nor a business with decades-old proven track of records.
Are people really going to buy ANT thinking it'll 10x to 50% of the US GDP?
Aren't there better stocks to YOLO on if you think you could maybe 10x? And if you want something safe instead of YOLO'ing, aren't there plenty of companies much safer and much more boring to invest in than something IPO'ing at 5% of the US GDP?
Isn't the entire Anthropic narrative based on infinite growth? How is infinite growth sustainable?
> Aren't there better stocks to YOLO on if you think you could maybe 10x?
This is why any [consumer] investment in ANT is bad. If ANT goes 10x then congrats, we [the citizens] are in a crisis. If ANT goes 1/10x then congrats, holders of ANT are in a crisis.
If we assume LLMs will continue to advance at a pretty rapid pace (relative to the current rate of advancement; so, same or faster), all these numbers look good to me.
Some people just start out with wanting LLMs/AI/Anthropic to fail or be scams and consistently try to fit any new evidence pegs into those unchanging hypothesis-shaped holes. This group ideologically despised LLMs from the start and needs them to be fake.
Lots of people making the financial case against Anthropic are not anti-LLM. I think the technology is great, interesting, powerful, innovative etc etc etc my issue is with the garbage business that Anthropic and OpenAI are building and the nonsensical ideas people have about business in general. Technology is exciting, business is not, business is boring. Money come in, money go out. Selling technology, no matter how exciting it is, does not change the nature of business.
And yet one can argue that the "garbage businesses" you label are the ones doing their best at matching market demand. See how well things are going for other entrants, like Google, Microsoft or Meta. Putting Copilot in every app makes a whole lot less sense than Cowork, Codex and the like. They just are doing worse things while attached to traditional money fountains, like ads in social media and search.
It doesn't matter how much they "continue to advance at a pretty rapid pace", if the economy has no actual demand for the new capabilities. Even if we accept the premise that LLMs can produce good code at scale now and automate most software development work, then what? As *an industry*, the bottleneck was that we needed people to write code, and that bottleneck has been removed since ~Opus 4.6. Where is the industry now bottlenecked which requires some significant quantum leap in model capability?
Opus 4.6's coding ability is nowhere near as good as Opus 5.5's. Some model next year will almost certainly be significantly better at coding (and planning, and designing, and ideating) than Opus 5.5.
You're narrowly focused on writing code. What about when they can run entire firms themselves, better than the average human CEO could? What about when we have models way, way, way better than Opus 4.6 and 5.5?
The point is we will eventually very likely have AI systems that can do everything. Everything. Anything and everything a human can do with - or even without - a computer, and then better than anything or everything a human can do, including in terms of creativity and innovation and novelty and insight. Obviously I don't know when that might possibly be, and it's not happening next year or something, but you can't just pretend like we hit the cap on the technology.
Don’t bother - these knuclkleheads don’t understand nuance and despite much of their devices being manufactured in China - they struggle to accept Chinese offerings can take significant market share.
>If we assume LLMs will continue to advance at a pretty rapid pace
Why assume that?
So much money have been given to that handful of companies and to that field, that it's pretty much guaranteed to plateau (it's not some small but growing market with lots of room to grow. It's already taking as investment a huge chunk of the economy, and the returns don't seem to justify it).
Most of it is a pre-IPO hype and a circular financing loop between Nvidia, Oracle, Amazon, and the AI players anyway, so the real returns are even worse.
You can be bullish on LLMs profoundly transforming society while being bearish on the economy "digesting" this new technology fast enough to justify the incredible capital spending necessary to power it.
No there are just numerate people pointing out ANT/OAI need to make X $$$ to make Y $$$ spend make sense. The financials are separate from model capabilities, they can cure cancer and invent big dick pills and still go out of business if their balance sheet is fucked.
Disagree on the tokenmaxxing end. People are still doing it (and I'd argue more people are doing it due to wider adoption). Its just that most low hanging coding adoption has reached saturation point very quickly (which is genuinely impressive).
Now, they need to unearth new use cases and saturate them to keep the momentum going. Its unclear if they have found these opportunities and can turn them into sustained growth.
Doesnt matter at this stage of growth what financial analysts say because they havent seen anything similar in their lives :) Large corps are going to really be the deciding factor, in what ANTs next years revenue is going to look like. And right now inside large corps there are literally hundreds to thousands of different experiments running and no one knows what is going to click.
That's not the takeaway of this analyst or anyone I have heard.
This is specifically about the end of tokenmaxxing and how much worse the growth rate would look after the Q3 numbers came out.
(I do think moats are shallow, and I personally switch away from Anthropic every few months for a while, but I've been coming back. All software moats have been made more shallow now, affecting the entire investment landscape.)
Anthropic is working hard to create a vertically integrated stack from model to harness to productivity apps. That may or may not work, but it’s clear they see models as a value add for integration and not a product that can be a moat in themselves.
harness is not a moat. CC is great, I mix it with codex with my own vibe coded messaging and hierarchy layer. Lately there are more CC because of Opus5.5 but Codex caught back up with Sol-6.1;
We have Copilot and can select between various OpenAI and Anthropic models, and when I accidentally select the “wrong” one it takes me awhile to notice.
I use Claude Opus at work and DeepSeek v4.1 Flash on my personal projects. For the kind of work I'm doing, they're basically interchangeable and DeepSeek costs 3% what Claude does.
Opus 5.5 is great, but you are kidding yourself if Deepseek 4.1 flash isn't as good as good for 99% of coding tasks. Anthropic (and OpenAI) are between a hammer and an anvil in terms of good enough local models being able to run performantly on current and next gen Apple hardware.
Does this person not own a calendar? Q3 numbers still polluted by expensive tokenmaxxing, which wind down late Q2 / early Q3 i.e. mid year budget review which will take rest of year to unwind. Q3 numbers = more mirage of tokenmaxxing tail and velocity = lagging indcator of tokenmaxxing end. Q3 also when fable tier coding commoditized and everyone doing a deepseek on cheap inference. The latter is what will kill ANT hardest, they signed up for 100s of billion of none cancellable compute contracts last year based on projected demand x expensive token cost, but demand / growth is less than token max timeline while inference cost getting cheaper. They have fixed, exorbitant costs, and operating margins negative for Q3, and Q3 just beginning of full price war. TLDR they will likely be vastly overpaying for compute. If turns out jevons isn't people using more expensive tokens on more expensive compute, but more cheap tokens on less compute, then ANT's entire balance sheet breaks.
>Q3 numbers still polluted by expensive tokenmaxxing, which wind down late Q2 / early Q3 i.e. mid year budget review which will take rest of year to unwind.
Evidence? I was told by a lot of people that token use was going to start to slow down 5 months ago because of "the end of tokenmaxing."
People PREDICTED token maxx was a fad/trend Q1 while it was underway, but companies asking employees to cut back was late summer, i.e. Q2/Q3 transition. And the argument isn't end of token maxxing its end of EXPENSIVE tokenmaxxing, i.e. tokens pushed at $$$ and volume that can make ANT numbers square. They're not precisely because people can now maxx on cheap commoditized tokens. The market for premium tokens (at least in SWE) got kneecapped, there's only so much demand for smart/fast.
Yes look at the dates... end of Q2, i.e. start of Q3, things take time to unwind. Marcus tweet MS cancelled contract, but doesn't say when terms end, but they almost always unwind over following monthS not ASAP. Real world works on a lag / budget cycles not how fast tweets / articles get published.
> Okay so Q3 ARR at $67.8bn versus $46.2bn for Q2, +47% QoQ. Turns out tokenmaxxing's end over the summer slowed revenue growth versus an absolutely torrid 100%+ growth rate in Q1/Q2 but 46% sequential is still eye-popping
There is an unbelievably enormous difference in 100% vs 47% -> and when you look at the 2nd derivative this is REALLYREALLYREALLY bad.
I mean, the growth is astounding and incredible. It does not bode particularly well for a ~$2T valuation.
We desperately need to ground these discussions in actual numbers.
This should help build some intuition and establish a baseline for what the $2T valuation means.
If you buy at $2T Anthropic needs to be worth about $3T by 2030 for a normal market return. That's ~25x P/E and ~$120B/year of profit.
That could be something like $400B of revenue at 30% net margins, $300B at 40%, or $600B at 20%.
It would take 1.7x revenue growth a year to get from today’s $68B run-rate to $400B by 2030. Anthropic's current growth rate is ~4.7x a year. Anthropic doesn't need to maintain that, they basically just need to prevent it from absolutely collapsing and hope margins hold up against open weight models.
You could look at this and say "yeah that's plausible" or "I highly doubt they achieve that" but it's hyperbolic to say its actually impossible they can meet this valuation, definitive proof of bubble, cant be worth this unless they can create God, etc.
> [Anthropic] basically just need to prevent it from absolutely collapsing and hope margins hold up against open weight models.
There are so many variables involved that are outside their control. If GPU/RAM demand dries up (or supply increases), for example, local models are suddenly a threat in addition to hosted open weight models.
The quality there is already good enough. I've been 100% local for personal projects since Qwen3.8 Flash dropped. It's not as nice an experience as Claude I use for work, but it gets the job done and I fully control my data and can do whatever I want with it.
You just made me wonder if part of the buy-up of RAM by OpenAI and other spend on GPUs and datacenters is literally trying to weaken the value proposition of open weight models.
Not just against open weight models and not just on hardware I think. Also on token subsidies etc. OpenAI and Anthropic both seem to be aiming to outspend everyone else, so that they can be the last ones standing. It is a risky move, especially since everyone else includes Google and Microsoft, players that have both existing surplus cashflow, solid warchests, and excellent distribution channels. But all outcomes are open still...
They are absolutely trying to deny their competitors necessary hardware, or at least drive up their costs substantially. And users running local models are, well, not not competitors ...
Interesting. Their revenue has "only" gone up 47% in Q3, compared to doubling revenue in most previous quarters. Meanwhile their "compute and infrastructure" cost has increased by 77% compared to Q2
Maybe they tried too hard to get good Q2 numbers, making revenue happen earlier and expenses later, and that now makes Q3 look worse? Either way, this does not look great
They clearly mention the cost increase in the amendment - it's because of the spot pricing of gpus (likely SpaceX). Their demand exceeded so much they had to quickly find some gpus at _very_ expensive prices
And they say that this will likely continue to grow until 2027 (when the long term commitment compute comes online, which comes with opposite problem .. will there still be enough demand)
It still doesn't look anything like a death knell. There is a point where you approach market saturation and the growth curve follows a sigmoid. Have you ever tried folding a piece of paper ten times?
I agree. Their financials are fine. They look like a company that has solid foundations, with room to grow but close enough to profitable that they can survive the end of the AI boom
But are they still a 1.5T USD company if their growth is already slowing this much?
So he can get all of his conclusions and predictions wrong based on his numbers and analysis, and then you think his value is in his numbers and analysis?
Funny how no one ever disputes his facts, only his predictions. He can easily be wrong on the predictions, while still being correct about the facts, i.e. their financials are shit and completely unsustainable. I don't care at all about his predictions, and am very interested in his fact-finding.
If you are trying to indicate that revenue is somehow not a key indicator of a successful business, or that somehow revenue growth is not the primary indicator used to derive Anthropic's eye-popping valuation for IPO, then I think you have lost the plot.
No, not when that business spends more than it makes to bring in that revenue.
A better way to phrase your thinking might be that revenue is an indicator of demand.
The issue here is that they've not really tested what happens to demand when the prices moves to where it needs to be to cover the cost of keeping the lights on.
>People should really look at cash flow statements instead of just earnings. Anthropic published one and their cash flow was positive for the first half of this year.
If you go just by operating income, they showed they can be profitable in Q2, and the operating loss in Q3 is fairly small (in relative terms. For every 32 dollars in revenue they lost one dollar). They haven't shown wild profits, but nobody expects them to. Any money they make has more value when it's reinvested into growing their market share and growing the market, making next quarter's numbers look better. And imho they have solid enough financials that they can make the bleeding stop if the money faucet turns off
All of that isn't very newsworthy though. We knew all that from the Q2 numbers. What is interesting is that reinvesting all their profits to make the number go up in the next quarter didn't work as well this time as it did all the other times
I'm unconvinced they have a moat that will allow them to ever show wild profits though. They are constantly 6 months ahead of open weight models. Their potential customers are perfectly able to wait them out. They'd need a government enforced cartel against open weights at this point for me to not view the IPO as a plain and simple cash out.
Some of my nephews ran a hugely successful pokemon card selling business (for ten year olds). I later found out that their mom was buying all their card packs for them to open and sell.
They stopped when the free packs stopped coming and they needed to finance their own.
There is no company on earth that can run on customers being interested alone, you also need to offer a product they can afford at a prices that generates profit.
Right now, I could start a business selling gold bars for $1 a piece. There would be more people interested in that than Anthropic.
Gross margins just means that they make more in revenue than their operational costs to offer the product (i.e., electricity, salaries, etc).
They have yet to release a model that has made more in revenue than it cost to develop...and each new generation costs more to develop than the preceding generation. Costs increases outpace revenue increases with every generation, across the industry, and LLM-based cash flows are highly negative even at Google and Microsoft.
Customer use of LLMs is decreasing after quarters of growth. Execs were all in on pretending LLMs were the future until the WSJ/Stanford report revealed what everyone outside of tech already knew: LLM use has not correlated with any measurable productivity growth in any non-programming industry. (And arguably, the same is true in tech as well: LOC has increased but quality has dropped, so the net change to productivity is de minimis.)
Since the start of the year the market has shifted from 50/50 split (closed v open weights) to 30/70 (70% of usage they see through their platform - is now open weight)
DeepSeek V4.1 Flash 1017 sees 5x the weekly token usage through OpenRouter than the closest closed weight model (GPT-6 Luna)
Even though OpenRouter is only one limited view of the market - those trends should be worrying to potential investors.
I'm not saying that OpenRouter isn't worth looking at but it, obviously, shows a much smaller slice of LLM market and it's going to, obviously, trend towards open models since those are the only model where there is any competition for inference pricing. No one is undercutting Luna because they can't really.
OpenRouter could so 90/10 split and it still wouldn't mean open models are beating (in usage) closed models.
Looking at the numbers there though, both Anthropic and Deepseek increased in terms of absolute usage numbers since the start of the year. And thats setting aside the high potential for sampling bias.
This view will ALWAYS overweight open models. The closed ones are being used in their own harnesses via direct API connections that OpenRouter will never have visibility to.
Let's say I'm starting a shitty cash-grab SaaS that needs some kind of "Ask Jenny" feature. Is the standard practice to sign up directly with a Claude/OpenAI API account?
The sentiment they are trying extraordinarily hard to create is that in the long run Anthropic will be the only one with any AI worth anything.
That makes any loss or risk reasonable.
It's great propaganda, so great that they have themselves started believing it with certainty. This is one of the self owns that brought down the USSR :)
Of course they could be correct too, time will tell.
Beyond it being propaganda, they're trying to work the regulatory capture angle to pressure lawmakers to pass safety regulations that would functionally ensure a market in which there are under a dozen players legally capable of competing.
And of course, obviously untrue to anybody paying any attention to the rapid progress in open weight models. The IPO will be a cash out for insiders and leave the typical institutional investors with the bag.
It could be untrue, but it's not obviously untrue.
The open weight models right now are nowhere close to ..say.. Fable 5.1
They could catch up if Anthropic reaches some sort of a plateau and they manage to keep improving, but there's no reason to be certain about that happening.
Google currently appears (by their latest numbers) to be sacrificing the only substantially profitable portion of their business to attempt to create that reality.
If we're talking web servers, sure. Video games? For the most part. Numerical high performance computing? ehhhhhhhhh i'm not sure history bears that out
Really? Even with this very generous 30 year gap ... you think you can run non-toy climate models on a $100 phone? Correlated quantum chemistry for a polyatomic system? Even just fully ray-traced rendering?
> You think we will eventually be able to run inference on a rasp pi?
Not today's raspberry pi, but the equivalent of it absolutely. I don't work in it, but have a number of friends who are involved in LLM development - the "breakthrough" optimizations are obvious work in hindsight. We are still in the very early innings of optimizing both hardware and software techniques. The recent market imbalance has blurred the idea that hardware simultaneously gets both cheaper and much better, but no one serious expects that to stop on a longer timeline.
Going beyond the "my watch has more computation power than NASA at the moon landing" - speech-to-text used to be unreliable tech that needed datacenters backing it as recently as the early iphone, and now runs in near-realtime on smartphones that cost about the same. A ~$50 processor running in my closet does realtime person/vehicle/package detection on my front door camera reliably.
Possibly. 1.58-bit models already run on CPUs https://github.com/microsoft/BitNet They're not good exactly but they are surprisingly better than you might think.
Meta? They're not even in the game. They don't provide inference to anyone. I think NVIDIA or any other hardware players that want to step up and take the bags of money awaiting them are the only actual winners. They also win when we inevitably move to local AI for many use-cases.
> Get rich following the bot's trade advice and never tell anyone about your bot.
The fun thing about open markets is the market will punish you for this with second order effects and it's in your best interest to get some whales on the bandwagon.
Leaving aside that Meta are selling compute. The reason I pick on Google + Meta is they are able to make their own customers + charge their own prices. Anthropic + OpenAI are both reliant on people coming up with Products powered by their AI models valuable enough to create several AWS worths of revenue and also not also not immediately moving it to open models.
Omitted/minimized: No mention of its models creating security issues in new code and the mass of vulnerabilities AI agents are creating in its wake. If you run Fable on Opus code, how many security issues do you find?
They were, but then Anthropic got so much flack in the business press for omitting them publicly (while sharing them privately with potential subscribers) that the bad PR was as bad as the financials themselves.
Yea, my guess is they decided they would rather have the actual Q3 numbers (and any reactions to them) rather than anxiety about unreleased Q3 numbers priced into IPO performance.
Which was probably the right move. The S-1 didn't seem to really change the overarching economic and financial narratives about Anthropic (it seems like whatever people already felt before the S-1, they felt it even more after looking at it).
So I don't think there's going to be any problem with boosters metabolizing the Q3 numbers.
Tokens themselves are already a bit of a commodity, and the margins are gonna be fierce on these as there is not a whole lot of lock in.
Would Anthropic and OpenAI position themselves using their extra tools, their ai agents, presentations, docs. Are they nice; yes. However, that doesn't make a 2T company. It reminds me a bit of Dropbox, a truly great product, however, they're differentiation is simply not there. I think the momentum is there for the IPO to go decent, but it feels like the company is on such unstable ground that the fallout could be catastrophic.
Agree. Was starting to look seriously into Google's AI stuff last night.
Owning the best search index and being able to retrieve content immediately straight out of their index seems like such a massive competitive advantage.
The market is completely divorced from fundamentals at this point. Just look at Musk's companies. Neither Space X nor Tesla have the fundamentals to justify even a fraction of their valuations, yet here we are. Anthropic wants in on the same bags.
In Q3, I think they struggled with the government ban on Mythos/Fable, Opus 5 not being so good, OpenAI models being more token efficient, and they decided to dedicate a higher percentage of compute to training.
The increase in token volume corresponds with a price cut. The S-1 revision already reflects what's going on financially. Anthropic is buying volume with margin. There's nothing wrong with that but we shouldn't expect profit to show back up as Anthropic's costs are increasing and have a fixed minimum. Given that OpenRouter users can switch models with the click of a button, what this demonstrates is that OpenRouter users are extremely price sensitive.
Why do they have to go with the ridiculous gazillion dollar valuations? Is it really such a bad thing to be valued at "only" 400B or whatever more realistic number, why does it have to be in the trillions?
the revised S-1 behind that $895B commitments figure says nine of the ten largest pharma cos, so it seems interesting that's what they seem to be basing off of
I have yet to see anyone releasing a nuanced report on all the metrics involved, everything I've found is short term or incomplete or misinformation, I am neutral, but the matter is complicated, they might pull it off, despite all the misleading marketing and Chinese competition and data center obstacles.
Too much sensationalist reports out there, cherry picked by all sides.
> It reported a $104 billion net loss during the period, though most of that was due to the change in fair value of convertible notes and derivative liabilities.
Look, I've seen and I understand the arguments that the GAAP income adjustment for this kind of thing is fake, but a number of them seemed to take it for granted that it was a "one-time" adjustment. If it's going to keep happening I really feel like we ought to insist on a more comprehensive argument. The accounting principle must be generally accepted for a reason: what shenanigans could someone get up to by reporting net income without this adjustment, and how do we know Anthropic is not getting up to those shenanigans?
It’s like shareholder dilution from issuing more stock or stock options. Existing shareholders might have a smaller share of the company than they thought. The accounting is weird.
We aren’t shareholders, though, so why should we care?
>We aren’t shareholders, though, so why should we care?
They're going through an IPO, so unless this is a one-off thing, people reading this thread will become shareholders, either directly or through their pension pots.
Sometimes accounting is weird because it's necessary to prevent tricks. Why do all your held-to-maturity bonds become "tainted" and get revalued at market prices when you sell a couple, even though nothing has changed about them? Because any other rule would let a company which cannot in fact hold bonds to maturity get up to tricks and systematically overstate their assets.
So while yes, this does look a lot like shareholder dilution, that doesn't really satisfy the question of why it's not treated that way under GAAP. Perhaps GAAP is just wrong, or perhaps there's some possible tricks which Anthropic is provably not engaging in. What it's definitely not is a one-time adjustment now that it's happened in two consecutive quarters.
It's a good question: why are the GAAP rules what they are? It needs to be investigated, though.
The reason it happens in multiple quarters is likely that Anthropic's valuation has increased in multiple quarters. Every time Anthropic's valuation increases, the convertible bonds are worth more money. Other shareholders have "lost" money compared to what would have happened if Anthropic had sold regular bonds instead of convertible bonds. Except that no, the value of their shares is increasing too. An increased valuation is actually good for them.
That's how equity works. If someone owns 1% of a company and its valuation doubles, their share is worth twice as much and so is yours and we wouldn't even normally consider it to be diluting the other shareholders.
So the question is whether convertible bonds are worth more than regular bonds and the answer is, yes, absolutely, but they are partially equity, and it's the equity part that's increasing when the company's valuation is increasing.
And if the company's valuation went down then the value of the equity part of the convertible bonds would decrease. Maybe that would be a profit by GAAP rules?
> The bond market works best when there’s an end in sight, Lyman noted. Every AI investor believes there’s a light at the end of the tunnel. We just have no idea how long this tunnel is.[0]
It is like someone who loses repeatedly at Roshambo while saying: "Okay, then. Best two out of three- three out of five- four out of seven- five out of nine..."
These are all supposed to convert to regular shares at or around IPO IIRC. If they're still pulling that crap after that point we'll know. There's an end date for those contracts that we're expecting
Only with money you can afford to have sit (hold) for 10+ years. The closest analogy we have dotcom bubble and crash. We don't know if we are in 1998 or 1999. For most investors who got in in 1999 and held, they didn't break even until 2007 or longer. Versus the investor that got in 1998, was break even in 2003.
Obviously the investors who bought in at the bottom did the best :-) Most people who are crying 'AI bubble' predicted 8 of the last 2 bubbles.
My personal opinion is ANT IPO is like the netscape IPO of 1995, its the start of the cycle not the end. Things are going to get really crazy over the next 3-5 years, the last year was just a warmup. I don't think people realize how much credit fiat dollars have been and can be created, and how $1T companies will be commonplace. I think humans have a mental block, where we think dollars are like rulers or other fixed measuring devices, so you can look at the same number at different periods of time and compare. credit backed currency completely warps our sense of scale when it comes to measuring things in dollars.
It's not the same thing as early NVidia / Amazon / Google IPO: these companies didn't IPO at a value immediately placing them in the top 10 (by companies market cap) and certainly didn't IPO at values representing 5% of the US GDP.
Can it x2? x3? x4? Maybe.
Can it x100 and represent five times the US GDP? I don't think so (I know, I know, the US GDP would grow too).
I know ANT ain't SPCX, but look at what happened so far to SPCX's valuation.
Half of the revenue in the first 9 months of 2026 was in Q3? No one is going to believe that. Focus on the net loss of $23.29 billion in Q3, that's more instructive.
Q3 numbers are not what you want to see heading into an IPO.
Again the issue is not is Anthropic an amazing growth story. Zero would debate that.
The issue is this is a probably $200-300B company pretending it’s a $2T company.
Add to that the insane amount of committed spend, a high interest rate environment which means they’ll pay a painful premium to patch cashflow gaps, and a likely “junk” bond rating (per Friday’s news) and that’s why folks are concerned.
It’s also just wrong to keep dismissing the paper losses from convertible notes as somehow not real losses. Once they are a public company this will be actual value taken away from existing public shareholders so they are legit losses not to be ignored. It’s like claiming equity based comp is free money… it’s not, you’re taking it out of the pockets of shareholders.
They may not be $2T company but they are also definitely not a $200-300B one. Their Q1-Q3 revenue is $33B with $17B in Q3. Even if they magically flatline for Q4, that's $50B revenue for the year. There are exceedingly few tech companies with P/S of 6 or less (equivalent to $200-300B valuation with $50B revenue) and we all know Anthropic's revenue will not be anything remotely like flatline for awhile.
> There are exceedingly few tech companies with P/S of 6 or less
I think an open question is if SOTA labs should be priced like software companies have been. Software has been so valuable because the marginal cost approaches zero as distribution increases. Profit/employee at a company like meta prior to massive AI spending was measured in the millions. Meanwhile, the balance sheets of the labs look entirely different - remaining competitive requires continuous and astronomic capital costs, and there's no indication that this goes away. They are perennially being chased by competitors offering the same performance as models mere months old at fractions of the cost.
Their business to me, at least long term, looks a lot more like an airline, railroad, or other extremely capital-heavy industry with very competitive and low margins. None of those trade at 6+ p/s. Particularly on the buyer side, the switching cost to me as a high-AI-spending employee if a cheaper model offers compelling performance is less than deciding that I should ditch my preferred business travel airline.
I think that's why you're seeing so many calls from the leaders of those companies to regulate their competition away.
But like Google and Facebook have a lot of hardware. They aren’t just pure in the cloud (though Google is running other peoples clouds). But yes their hardware demands per dollar of revenue is much lower
Ultimately it's marginal, meaning that every $ of revenue that Google and Meta make have an associated compute cost with it, but scale your revenue down and you scale compute down also. Moreover this compute cost per revenue is fairly low. A small cloud instance equivalent's of compute can generate a ton of business value.
The big foundational labs are not like this. Every model they release takes a ton of data infrastructure and training (pre, train, post) compute. Then inference itself requires compute. Even if the labs sell tokens, each token embodies a lot of training and inference compute. This is a much more capital heavy business, so I can see revenue multiples being smaller than pure SaaS businesses. A model release bust (like Opus 5) ends up leading to big losses because of the necessary training compute spent on the model and the labor spent on training. A bust SaaS product ends up being wound down quickly, and only costs the labor spent on building it and the compute spent running it.
A good analogy is chip companies before all of the AI related compute pressure. These were largely low margin businesses because of the fixed capital cost associated with each new chip gen, architecture, process, etc
Well, it's true that Google has a lab-like structure. Computing the index is a massive continuous "training" job that takes multiple entire clusters (when I worked there at least), and a Google cluster could easily consume an entire rented colo facility. And you need to have crawled large chunks of the web, then done a lot of expensive calculations, before you can serve a single query. Scaling down the crawl/index pipeline is possible to some extent but not much - no different to LLM training.
So I think the reason for the value gap is purely in the proprietary IP. The recipe for making a good LLM is more or less public, unless you buy into the "everyone just distils claude/gpt" angle. The recipe for making a good search engine is private.
The value in Facebook is in the social network that can't be replicated by just buying servers. The value in Google is/was in the secret IP. Hardware costs are low compared to the value of those. Hardly anyone knows how to make a search engine that works as well as Google does, or how to make a social network that draws everyone away from Insta/FB (ok bytedance do...).
Note the huge difference in strategy and outcome!
Google published essentially nothing on how to build a search engine. Huge R&D spend, zero publications, intense culture of secrecy, extremely low attrition to startups or competitors. They did publish some papers on big data systems but only on tech that was many years old - a typical Google "research" paper pre-2015 would announce that the system it described had been running at scale in production for years already. Google also never faced any kind of split or mutiny in which lots of people with the core secrets left the firm to set up a competitor.
With AI however this changed and they published all their research immediately, also their best employees constantly ran away to join startups. The result was predictable: smaller competitors deployed their ideas faster and better than they did, and now Google has no moat anymore.
OpenAI took the classic Google approach and stopped publishing R&D after developing GPT but this didn't help them build margins because:
1. Apparently the core transformer idea is powerful enough that breakthroughs after that didn't matter much. Nobody has been able to develop some killer edge that leaves competitors unable to catch up. Most improvements to transformers have been about cost reduction, not capability improvement.
2. They had a huge split when the Anthropic guys walked out over safety, taking the knowledge of how to build a frontier LLM with them. Result was a peer competitor from day one, whereas Google had at least six years to build edge before anyone realized web search mattered.
Result: Google-style hardware costs but without any secret IP that could generate revenue commensurate with those costs.
The scales are so dramatically different that it's akin to pointing out a home office and a steel plant both consume electricity.
> we all know Anthropic's revenue will not be anything remotely like flatline for awhile
The SOTA US labs are one chinese open model away from being abandoned overnight by the masses. Ive argued before, but argue again here, that the current hardware shortages are the main thing keeping this from being a reality currently. If more people were able to run local models for a reasonable all-in cost, we would have a much higher velocity of iteration and improvement in the open source community. But its still too niche while people cant justify the cost, and so they continue to pay the premium for inference.
The limiting factor now is more electricity than hardware. There is hardware sitting idle because data centers can't get grid connections and power delivery contracts. Even for those who willing to run gas turbine generators on site there are severe supply constraints.
I dont think electricity cost is what is keeping John Doe the developer from running local models.
OK, but that's irrelevant to Anthropic's IPO. For the next few years at least very few developers are going to do much with local models. The problem that new data center construction projects are running into is electricity availability. In some areas there's no more power available regardless of cost.
You should stop arguing that because it doesn’t make any sense.
If a Chinese company released an open weight model that somehow was so superior to Anthropic and OpenAI then they can just take the model and deploy it, reverse engineer it, use it to improve their models, and/or do any number of things they currently do today which are of increasing value.
Most people don’t want to or have any interest in running or managing models locally.its not just about cost. The market for running models is infinitesimal in comparison to the market for managed products. That’s why everyone uses Office and not Open Office. Nobody is going to run some Chinese open weight model no matter how good it is on their critical banking infrastructure or something.
There will be space for lots of different solutions and many products and types of models including Chinese open-weight models and American ones too once we get around to making money on them.
The game isn’t just “cheapest open-weight model at all costs” which is what you are asserting here.
> If a Chinese company released an open weight model that somehow was so superior to Anthropic and OpenAI then they can just take the model and deploy it, reverse engineer it, ...
> The game isn’t just “cheapest open-weight model at all costs” which is what you are asserting here.
Decide: is the claim that open weight models will be more intelligent than Anthropic/OpenAI SOTA models, or that they will be cheaper?
> Most people don’t want to or have any interest in running or managing models locally
Most people are also not the primary customers of Anthropic! 5 companies make up 25% of Anthropic's revenue, and I bet there a ton of companies spending tens of $M/month on Anthropic/OpenAI that make up much of the rest. These are all medium/big companies that have the means for self-hosting models if they decide that's more convenient. The cost is of course not the whole story here, since Anthropic/OpenAI generally make better models and that allows using them in more situations, where perhaps other models would fail, but this is also a big cost for Anthropic/OpenAI to continue getting ahead, and the moment they stop they lose this advantage. Also in my experience self hosted open weight models are _fast_, which is _very_ nice for all those tasks where you don't need SOTA intelligence to get things done.
Every country in the world outside Europe and United States will adopt/use an open Chinese model because it will cost them relatively nothing, the only thing they have to supply is hardware. And the Chinese certainly can provide that. And so can several other companies outside of China. But those companies won’t be OpenAI, Anthropic or even Nvidia.
Where is their moat? Ultimately, where is their profit? answer? There is none at least there is none that will come close to whatever imaginary valuation Wall Street comes up with, in the best of times, while Wall Street may be able to carry one such company, but two within the same year?
High profit software products have monopoly pricing power from network affects: everyone uses office for compatibility/familiarity with what everyone else uses, everyone uses Facebook because everyone else does, everyone buys advertising buys from Google because that's who everyone sells to because that's who everyone buys from.
No one has shown a network effect for models yet. Models are commodities. Anthropic etc will make the sort of profits one makes selling commodity goods (a couple of percent), plus a perhaps a bit of extra margin from being a few months ahead of open weight models (if that profit can justify the billions in training costs).
I do agree that local models are irrelevant. But local inference and open weights are different things. You can go on openrouter and see that there are dozens of companies selling open weight inference.
The important distinction is SOTA US labs - ANT and OAI - who can not economically survive reverse engineering / fast following PRC models at PRC pricing. Like obviously PRC operators are not going to flourish in US due to geopolitics alone, but that doesn't stop US operators running PRC models, or reverse engineered Murican open models from creaming US SOTA labs on economics eventually. The closer open models are to SOTA in capability the smaller SOTA TAM gets, and that market might not be large enough to fund/maintain SOTA. SWE is one of the highest margin sectors to capture, and open models already shrunk SOTA niche to critical infra and cybersecurity etc then what other high $$$ niche market going to replace? BioPharma maybe? The point is SOTA capitalization depends on large capability gaps that economically justifies 30T TAM - the more domains open models fast follows to compress TAM, less places SOTA has to carve sufficiently profitable moat, and less sustainable maintaining SOTA becomes, especially for dedicated AI labs with no alternative revenue. It's not cheapest open-weight wins, it's cheap and good enough open weight simply murders current SOTA economics because SOTA likely cannot finance at all costs anymore.
The SOTA can't exist in this model because no company or investor will pay for something that will just be undercut.
What you are arguing is that all progress on models (outside of government labs or internal corporate workings) will effectively stop because there won't be funding for it. That seems unlikely to me. You just need to play it out...
US companies/investors go and invest a trillion dollars or something on models. Someone "does stuff" and reverse engineers their model. They go out of business (you said so [1]).
So now what happens?
Well, the US/DARPA/universities/whatever (all countries with interest) will keep working on this stuff and there will be some funding for it but that's about it. Military applications, physics, things like that.
Chinese companies will also fail and it will not be possible to create new SOTA ones because they'll obviously just be reverse engineered too and given away for free or at super low pricing - whatever happens in America must happen in China too.
So who will build new models? There's nobody around to pay for it besides governments or maybe corporate labs. But corporations and governments will protect the models (obviously, as OpenAI and Anthropic do today to varying extents) and keep the weights and stuff closed source. They'll have some productive use for them even if it's not in the download Claude Code and have at it style of B2C/B2B we have today.
Or
Maybe the trillions of dollars of investment going into these companies actually makes sense and the threat of open-weight models isn't what you and others seem to be suggesting it is. Whatever fate we seem to want to assign to America also happens to China here. To the extent they are behind, they're still running the same exact race. If open-weight models undercut companies they'll also undercut Chinese companies.
Just as Open Office exists today, so does Microsoft Office. Microsoft just makes all the money. I'm not sure why some seem to think that they can use the open source comparison here but not also use the open source economics that we've seen play out.
I think as a productive exercise someone should run this argument and explain why Chinese companies won't also be eaten alive by this alleged race to the bottom and have all of their stuff reverse engineered and given away. How can they invest in new AI companies with SOTA aspirations? Whatever argument you want to make for the US applies to China as well, and we can just give away freedom models to all the countries just like supposedly they will.
[1] "The important distinction is SOTA US labs - ANT and OAI - who can not economically survive reverse engineering / fast following PRC models at PRC pricing."
If SOTA finance model is currently unsustainable, then it's entirely economically plausible it will stop. US investors and SV pouring 100s of billions in accumulated $$$ from decade of cheap money printer and tech war chests into one massive hedge is anomalous business model. Did it make sense of TAM is 30T... maybe, but is TAM still 30T? We are seeing currently most profitable SWE niche get creamed, and other sectors likely follow.
>whatever happens in America must happen in China too.
No, because PRC literally could not do what US did... they simply didn't have access to enough compute to dump so much $$$ capitalization for SOTA race. On the other hand, on paper US WITH MORE compute and (currently) CHEAPER compute can put forth cheaper models to undercut PRC open models... but functionally they can't for the same reason cheap PRC operators can't win in US - geopolitics. Undermining PRC AI requires giving PRC cheap US AI on cheap US compute which is sanctioned. But ultimately that is entirely irrelevant to the fact that CHEAPER US models will undercut EXPENSIVE SOTA US models. The undercutting is not symmetrical. US companies can use cheap PRC hammers (or even cheaper US hammers) to break $$$ US SOTA. US who on paper can provide even cheaper hammers to PRC... but cannot sell those hammers... so PRC stuck using cheap hammers... and since PRC never had chance to spend $$$ on SOTA, there's no equivalent valuable things to break even if they had cheaper hammers.
> trillions of dollars of investment going into these companies actually makes sense
We know the context for which those $$$ investment makes sense... if SOTA held meaningful lead to sell high volume of $$$ tokens which was hedge 1-2+ years ago when years of $$$ commitments were made. But doesn't seem to hold anymore, ergo it no longer makes sense. Microsoft had decades to build ecosystem and moat, just like adobe it was, actually hard and high friction to move away. People are switching models all the time. SOTA AI only labs has much less moat than Microsoft or other incumbent tech with decades of overlapping capture mechanisms.
> I think as a productive exercise someone should run this argument and explain why Chinese companies won't also be eaten alive
The productive answer is THEY ARE ALSO BEING eaten alive.
The even more productive answer is PRC commodity AI doesn't have to win for US SOTA AI to lose.
There's just not much meat on PRC AI bones so they are losing billions instead 100s of billions / trillions. Like there is ongoing domestic AI involution irrespective of US SOTA. Literally the primary reason PRC AI was forced to go free was because DeepSeek is billionaire side project who priced his hobby as free which set economics for all other players. Being sanctioned/compute constrained = they figured out how to close SOTA gap on the cheap but also set ceiling on how much they can lose on race to bottom. In so far their SOTA aspirations, again refer to DeepSeek, he takes some couch change from hedge fund gains or investors, and throws it at smart people and hope it's enough. If its not then they'll probably stop too. What PRC not going to do is mobilize resources equivalent to several % of GDP concentrated in few entities for SOTA AI, at least not until their compute is sorted, in which case competitive exposure / risks becomes symmetrical. In the mean time, things are not symmetrical and arguments against economics of US SOTA AI does not apply to PRC commodity AI.
> No, because PRC literally could not do what US did...
I think the fundamental disagreement is you believe that cheaper models can just break the state of the art. This seems false because if it were true then the SOTA can also stop developing and spending on SOTA and just use a cheaper model to build their own better models and leverage their existing compute and energy infrastructure (or even build it out for better and more capabilities later). Anything that is released the SOTA has access to, but it also has access to things that others don't. It's ahead by definition and it's not possible to come out ahead of it because the SOTA will just absorb whatever open-weight/open-source/whatever is released and add to its own lead.
> SOTA AI only labs has much less moat than Microsoft or other incumbent tech with decades of overlapping capture mechanisms.
And those companies are worth a lot of money, continue to execute at high levels, are systematically important, and are by any definition wildly successful. That's my point. You have to apply the open source style argument fairly and equally. If you're going to suggest the cheaper models (which follow this open source software argument) are going to come into being and do all this stuff you also have to acknowledge that they made no money in comparison to the Microsofts of the world and Microsoft won because it made all the money. Whatever open-weight stuff that gets released everybody gets to use it and so tech companies get to use it too and accelerate their advantages. This is what we've seen play out, why would AI be any different when you're using the same economic paradigm as your argument?
> Literally the primary reason PRC AI was forced to go free was because DeepSeek is billionaire side project who priced his hobby as free which set economics for all other players.
American billionaires can also "just set up a side project" and reset whatever economics exist. The crux of your argument is someone can do this and then SOTA isn't economically viable, but you can't escape that this has to be true for everyone, not just the US. If it's true for everyone you also then have to admit the SOTA can't progress because there's no viable economic model - anytime you make an advance in the SOTA someone "just sets up a side project" and/or steals your model or w/e.
The US SOTA losing just means the SOTA has to be equal for everyone or the US is ahead - you can't have it both ways. Progress is practically impossible because you can't make money and you can't obtain a lead so nobody will do any work since their work will just be immediately utilized by someone else.
If what you are proposing here is true, then China is permanently behind. There isn't an economic model to obtain a "lead", yet the American companies get to just take all the work that Chinese companies do (until they go bankrupt not making money) and add it to their lead.
> The productive answer is THEY ARE ALSO BEING eaten alive.
Well you can just run them out of business then. If they spend any money you just spend more until they can't make money and close up shop. Or is someone going to subsidize them indefinitely?
The economics of these arguments that the US SOTA companies are just going to go bankrupt because of cheaper models is something I just don't find compelling. I've been trying. I debate these things so I can learn, but I'm just not seeing a coherent argument on the topic.
I don't think all of these tech companies and investors are stupid. Greedy to a large extent and some will lose their hats? Absolutely. Stupid? I don't think so. Worst case is all the infra is still useful.
> I think the fundamental disagreement is you believe that cheaper models can just break the state of the art.
No their argument is that the existence of a worse quality cheaper product will eat too much of the market for the SOTA company, making the SOTA company less profitable (not necessarily unprofitable) because the number of people who need/want the quality difference supplied by the SOTA product aren't enough to pay to sustain the extra resources required to supply the next iteration of the better product or even if they are then they are not enough to get a respectable ROI on the current investment.
>SOTA can also stop developing and spending on SOTA and just use a cheaper model
Yes, and they will be forced to (if the market shapes out as above) but the point is that then the company won't be getting the returns that justify the valuation being talked about in the IPO and will likely never return the amount of money that has already been invested in it even if it is a profitable company. i.e. If I spend $20,000,000 making a lemonade stand that is a successful best in class stand and it nets me $50 profit a day then it doesn't matter how that as far as lemonade stands go that it's a pretty amazing example. I am effectively never going to see that 20 million again let alone a real profit.
They're not arguing that these SOTA companies aren't better, they're arguing that there won't be a decent return on investment let alone justifying further investment and means everyone will be taking a huge haircut on what is already invested (with the possible exception of early investors who take the bag with the IPO when it inevitably succeeds based on people disagreeing with this evaluation and thinking it a good investment or simply because of index funds being required to buy in)
>has to be true for everyone
Cheaper models break financing SOTA pace, PRC budget fast follow currently is catching up. If SOTA slows/stalls slower SOTA pace = instead of PRC 6 months behind, they're 3 months behind, or period of overlapping parity before US SOTA grabs lead again for a few months. There may not be economic model to maintain durable lead, which SOTA economic doom spiral condition. Now if PRC fast follow couldn't close gap, i.e if PRC only fast following due to distillation and US SOTA can technically prevent PRC fast follow, and extend gap then maybe exorbitant SOTA spend sustainable. If not sustainable everyone moves to cheaper cycles. This an issue if SOTA cannot be done on the cheap, which also true for PRC, but PRC not doing expensive SOTA, they're doing cheap fast follow.
>why would AI be any different
It's not, someone going to grab big share of AI... but that someone has to be competitive enough for exquisite lock in, and frankly with enterprise reducing SOTA AI spend due to SOTA AI costs, current SOTA AI ain't it. See last paragraph for hopium.
> someone going to subsidize them indefinitely
Yes. Same way Alphabet/Meta can subsidize AI, China has their Baba and tech giants with alternate revenue streams to subsidize AI, even if not SOTA, and geopolitics will make sure US cannot run PRC AI out of business in PRC and vice versa. But geopolitics not stopping PRC companies running other PRC companies out of business, same with US companies running US companies out of business. In this case host countries tech giants with high margin businesses squeezing sole play AI labs. It's just US SOTA ANT/OAI getting murdered is massive economic contagion risk for US vs Zai/kimi not for PRC.
> find compelling
This really not complicated, last couple years, US SOTA labs signed 100s of billions of compute based on projected demand that they could monopolize enough demand to finance compute capitalization. This not stupid investor prior, because their only competitor, PRC was export controlled, compute constrained and shouldn't close gap. But PRC closed gap, made models good enough and architecture efficient / cheap enough so other players, including those in US, can capture large % of demand WHILE using fraction of compute. SOTA labs now looking at margin compression, losing share to good enough competitors and obligatory (now excessive) capitalization. This bad combo for current US SOTA economics. The numbers suggest this is what is happening. There is chance demand increases past projections and extra SOTA compute will be boon, but again, we are seeing most profitable SWE enterprise curtail not increase demand.
> Worst case is all the infra is still useful
IMO this actually best case for US medium term favour, but also another strike against CURRENT US SOTA. Reality is current SOTA is grossly overpaying for compute / AI premium. Even if OAI or ANT dies, their compute can be picked up for pennies on dollar and as long as PRC behind on compute there is potential for cheaper future US SOTA economics from cheaper hardware since crash will lead hardware price regress towards mean. But that boon is for the next guy.
Here's the US hopium: post current US SOTA crash, there is potentially curve where US has so much cheap discount compute, because current SOTA taking the haircut, and where PRC has not much compute, that US will have sustainable SOTA prospects, i.e. the world where US pays $1 for shovels instead of $10 and can buy $10000 worth of shovels while PRC can only buy $100. That's feasible after current exorbitant SOTA economics bleeds it's patients to death and the next US SOTA host get the transplants and lighter balance sheet, i.e. web 2.0 for SOTA AI. Incidentally more likely "winner" of AI... if I had to guess assuming incumbent FAANG doesn't just snatch up ANT/OAI if they explode - new player/entrant who are near SOTA and eventually SOTA but is priced more competitive than AI 1.0 SOTA, that enterprise will gladly pay for since not at current exorbitant rates that's causing enterprise to bail, but it will be fine because AI 2.0 SOTA can be sustained from 3T TAM instead of 30T TAM. Now PRC might still be able to fast follow, but US SOTA economics completely different if cheap hardware = US premium SOTA cost only 2x more than commodity PRC tokens instead of 20x.
> Most people don’t want to or have any interest in running or managing models locally
When I argue the hardware cost, I should note that I think this is the limiting factor for _improvement_ of these models. But ultimately I dont think most users will self host in that world. They will still pay for inference, but it will be for one of a dozen commoditized vendors who are able to reap the benefits of inference at scale.
It could happen, but it would be an anomaly. In general if a business is to make money it does not let itself become a commodity vendor. You can't go to General Burger and select from one of a Wendy's or Big Mac or Burger King burger. You can't go to Microsoft and buy a Playstation. There are exceptions, of course, but whoever is building this stuff is going to want to make money and becoming a commodity vendor isn't the path toward that goal.
It's the same reason that open source software has been a commercial failure except for a few evangelists or whoever becomes a cool "thought leader" by releasing a new library and then gets invited to conferences to speak. Software was commoditized in many ways, yet all the big software vendors made more money than GOD because they took whatever commodity parts came to exist and monetized them.
I think the models can be improved with either recursive improvement (it would be faster with better hardware but if you can do it at all you just need time) and algorithmic developments. I don't know what those would be, but usually we see improvements to both over time. The algorithms will tend to be trade secrets as well.
Sure, in an ideal Chinese-world, Chinese companies would not be open sourcing their models and weights. But I think its very clearly a deliberate effort to cut off the US-companies at the knees by doing so, as they are unable to compete on the marketplace directly due to the current business/political climate. I also dont think these need to be better than SOTA, just reach a "good enough" bar. Big businesses will continue to use the best-of-the-best, but that will not be the general public and also will not justify current financial outlooks (in my opinion).
Outside of the United States, Europe, Australia, New Zealand, Canada, the rest of the world can choose their computer and choose their model, and they will.
I don't think US companies are being cut off at the knees though. I don't use a Chinese model yet I pay a monthly subscription for Claude. Is there a cheaper Chinese model? Maybe? I don't even care enough to be bothered to look. There are tens of millions of people like me who are in the same boat. Plus you also just have high-end research (as does China) which won't be permitted to leave the country.
The problem here is that let's say it's true that the Chinese are cutting off US companies at the knees. Well the US can just do the same thing back to the Chinese. We might have some money lost and stuff but the data centers will continue to be useful, as will all the infrastructure and energy grid buildout so it's not a total loss (I couldn't care less if accredited investors like myself lose money - you know the risk). Then we just steal Chinese models, do the open-weights thing, and all that stuff. So where does that leave us? Nobody makes any money? (Also where is Europe in all this?)
> I also dont think these need to be better than SOTA, just reach a "good enough" bar.
I agree with you in some respects but I'd like to digress here a bit. Good enough for a general purpose model isn't good enough. You want the best or at least what performs to your expectations. I don't even care about old models nor how much they cost (in my personal life and of which I will be a paying customer as will tens of millions of other people) I just always use the latest and best one unless something else triggers me to use a different one. It reminds me a bit about software updates. Folks around here are like, I'm not upgrading to macOS 17.2 and I'm staying on 17.3! Regular people? Not only can they not tell the difference they don't care - just give me the latest and best thing. period.
On the other hand you can have models that are OK at general purpose tasks but are perhaps specialized for certain applications (software or hardware). Those could wind up being on a spectrum from cheaper (and integrated with lower-quality products kind of like an Android Tv model) to very expensive and high-performant.
I don't see a world in which people are by and large sifting through well today I want to use Kimi 1.1 and well later today I'm going to switch over to Claude Lite 18.7 and dang it's not giving me what I want so I'll switch to whatever model. That's not how people operate - they don't want the cognitive overhead of trying to mess around with something that's good enough. They may have to suffer that based on their economic means, but it isn't going to be an expressed preference outside of very niche categories (like home-brew hackers playing with their toys).
Similarly this is just my opinion as well. I enjoy a good discussion here and I hope you are as well. Positive intents, even if my words my be curt, at times.
> Similarly this is just my opinion as well. I enjoy a good discussion here and I hope you are as well. Positive intents, even if my words my be curt, at times.
I have no issues with this thread : D But back to the discussion...
Again, I elided a lot of thoughts on the whole situation, but the other factor in my belief that this is an inevitability is that the SOTA labs churn too fast in both model behavior but also harness behavior. Normie businesses of the future are not going to hitch their wagon to a moving-target of both behavior and enforcement vs just pinning the equivalent of "opus 4.5" from a run of the mill inference provider. It just doesnt make business sense to be at the whim of a model update. And the entire perpetual motion machine aspect of the SOTA financial outlook _necessitates_ constant churn. If they dont release, they die.
> So where does that leave us? Nobody makes any money?
I think this leaves us with SOTA labs pushing hard for protectionist regulation to essentially require federal blessing to be an inference provider. I also think we will see a _ton_ of dark money poured into research to try and show things like theoretical smuggling/side-channel exfil attacks on models/model weights so that the SOTA companies can paint the _open models and model weights themselves_ as a bogeyman. If they can show that theres potentially latent attack vectors in those paths then that will then prevent all the "middle class" companies from being able to use them, especially if they need to adhere to eg SOC2/HIPAA certification etc
Worldwide, it is will be the cheapest model. That is going to win out. In the United States. and Europe, that will be their (OpenAI/Anthropic’s) markets, every place else probably will be a Chinese Open solution. And that will be no different than EV cars. or solar panels the Chinese dominate worldwide and the same probably will happen with AI models outside the United States, Europe, Australia, New Zealand and one or two other countries.
Also, because of the three hyenas of memory, the Chinese probably eventually will have control of the memory market too in the near 3 to 5 year future.
> Worldwide, it is will be the cheapest model. That is going to win out.
Win on what terms, or in what sense? We've seen this play out via Android versus iOS for example. The cheapest models are more prevalent, but Apple makes all the money. I'd be perfectly happy being the Apple in this case.
It's like you can go and buy a cheap Android phone. It technically does all the same things as the latest iPhone, but it does them worse. So Apple ends up making all the money and delivering the product that people go on to use to do all sorts of other economically advantageous things that make them even more money. Where does that leave the Android phone/model? In the dust, for the most part.
You are misunderstanding a number of things aggressively.
For one thing, one needn't run a model locally, that's a separate--but also very damning--issue.
For another, that's fine if Anthropic and OpenAI can do that, but they will have sunk a bunch of cost into their own training already and people wouldn't pay them what they need when there is always a vastly cheaper substitute that gets the job done.
> we all know Anthropic's revenue will not be anything remotely like flatline for awhile.
How can you confidently say that? At some point they oversaturate and growth stalls. Didn’t F100 have a major pullback already this year on AI spend?
This is what makes a market. If they keep growing those waiting for the bear case will be left behind. Those who are on the side that the biggest most important AI company will continue to grow revenue will win.
"Biggest" and "most important" are not given, settled facts
Isn’t the whole reason companies go public now to dump on retail? What growth is left?
SpaceX is down, Cerebras is down, etc.
Anthropic has something close to first-mover advantage but that doesn't mean they'll continue to grow revenue. Ten years from now the dominant player could be Google or DeepSeek or some random startup that doesn't even exist yet.
I didn't know that the "left behind" rhetoric also extends to investing into their stock?
I guess it applies to everything. Universal. Like LLMs.
> This is what makes a market.
What is "this" referring to here?
https://news.ycombinator.com/item?id=50339625
the difference in which side of this bet people take.
of course each stock price is many bets intersecting at the same time (with all their correlations of course)
stocks are claims on future cash flows, currency is claim on future purchasing power, but all can be though of as forms of money, where the prices are based on background information [ie. priors, beliefs, experience] and people's prediction of the future supply and demand.
and the general abstract definition of said bet is asking where are you in the information flow. if you know with sufficient certainty how the people afer you will evaluate some news then you can guess which side to take to make some money.
The only justification for extremely high P/S ratios is if there’s strong evidence that the future margins and profit will be amazing.
With the price of AI rapidly falling to the price of compute, with models therefore basically having zero value, and with Anthropic having rented all its compute from others then there’s little evidence that such high profits are in the future.
Look at the Supermicro, their expected revenue is ~70B for FY2027 (was 39B in FY2026), and they are valued at 20B, and they even have some profits.
200-300B might be plausible depending on how market perceives the company
It's not that hard to get 33B revenue with the 104B net loss.
Anthropic has a great product but is for sure a bad investment.
I wouldn't be so sure as to say definitely.
There's a world where capabilities narrow and Anthropic is stuck selling a commodity (not like they don't have any competition now). In this world it's not unbelievable that Anthropic would trade around or below 1x revenue no?
I think calling Anthropic a tech company could be like calling an electric utility a tech company.
I think this is focused on the wrong side of the distribution. You can view the value of bear cases at ~zero and it doesn't move the math much.
The $2T comes from the implied odds (perhaps 10 or 20%) that we get near term AGI or RSI and AI becomes the entire economy, crossed with the odds that Anthropic gets there first and is the runaway winner (maybe 25%).
The trouble, in my view, is that today's revenue numbers are essentially irrelevant to the outcome that drives ~all of the expected value. The market has been pretending otherwise, but fundamentally we're looking at a globally top-10 public company that boils down to a binary call option, and the probability underlying that option is inscrutable. That's never happened before, and an inability to divine a fair valuation properly demands extreme volatility, especially in reaction to capabilities progress.
Finally (total speculation): I think Anthropic knows this. I think they're sitting on something with Fable 5.5 and the surprise outperformance of Opus 5.5 is a tell. They intend to release it just before the IPO and make some big claims underwritten by its performance.
> globally top-10 public company that boils down to a binary call option, and the probability underlying that option is inscrutable. That's never happened before
FWIW Bitcoin had the same dynamics. Its valuation between 2011-2021 basically derived from a superposition of two futures: one in which it took over the entire global economy and became the dominant reserve currency, and one in which it failed and its value went to zero. The middle-ground was effectively excluded because currencies have strong network effects and a currency that is not dominant within a market quickly ends up unacceptable to merchants and worthless.
It's still somewhat in that grey zone, used as a form of digital gold by people who believe that the dominant reserve currency (USD) is going to zero as well as a trading medium by rogue nations, but its performance has been lackluster since 2021. It's currently up about 20% since then.
Entirely possible that AI follows a similar trajectory. People assume that it's either going to go bust or take over the economy, but people are really good at overlooking excluded middle grounds. Personally I think it'll settle into a space where it's very effective as a coding assistant, effective at creating propaganda and fake forum posts, and has the effect of killing Google Search (but not really replacing it), but doesn't fulfill all the hype around completely replacing the economy.
> The $2T comes from the implied odds (perhaps 10 or 20%) that we get near term AGI or RSI and AI becomes the entire economy
At this point every churches should IPO and raise hundreds of billions from the implied odds that judgment day is near and Paradise becomes the entire economy, that wouldn't be crazier.
many cults do this regularly
Anthropic simply has a better claim, better PR, more testimonies, better prophets :)
most people have been already inoculated against that kind of doomsday predictions
(and also because of this there's a huge problem with solving slow moving diasters, like climate change)
Good points.
In addition to:
> the probability underlying that option is inscrutable
The "positive" side outcomes are also pretty inscrutable imo: even if AI completely replaced labour, how does it trade against natural resources? Even if anthropic gets there first, might open weights arrive a few months later? If AI drives the value of labour to cents per hour, does that crater global demand?
For sure, the "anthropic stock is a ticket to heaven" scenario is in there somewhere, but even taking complete AGI victory as a given, it seems like a very narrow outcome.
On the third hand: if AGI victory is assumed, even if Anthropic isn't a great bet even under that assumption, where else to put your money? While the music plays, you have to dance.
Another consideration: if we get AGI and it becomes the entire economy, what are the odds that the money you've invested still matters? It seems like both the "universal post-scarcity" or "AI company controls all wealth and power, the peasants live in squalor" scenarios are likely outcomes.
Aren't their operating expenses quite high? That is they depend on high expenses to continuously drive revenue. If they can't keep raising money, then their revenues should be expected go down.
And what are the debt levels of these other tech companies?
Rumors going around that they want to ipo at $1.5T. If true there is an argument to be made that they are cheap.
> There are exceedingly few tech companies with P/S of 6 or less (equivalent to $200-300B valuation with $50B revenue)
But a big part of the reason this has been true is that (mostly-software-based) tech companies haven't historically been spending significant billions of dollars every year in capex.
The other tech companies have such a thin and easy to vaporize moat too?
Because we're just waiting for the world to find out that they can spend 1/10th and get virtually the same results.
Open source models are now 1/100th in price per task of US sota from this summer...
Before you start using multiples - you should know that p/s multiples are internally inconsistent.
Well, it depends on how you interpret the drop to a 47% quarterly growth rate.
If this is the new normal, that's great news. Revenue growing at 47% per quarter will 10x in 6 quarters. Even if you think it should be a $200B company today, it'll be a $2T company in 18 months, assuming the same revenue multiple.
If you think this is the start of a rapid decline to 0% or negative growth rate, then it's obviously bad news.
The real answer is probably somewhere in the middle -- but exactly where in the middle you think it will fall depends more on your own priors than anything you can read from the data.
> Once they are a public company this will be actual value taken away from existing public shareholders so they are legit losses not to be ignored.
They reported a loss here because they sold notes convertible into shares and their valuation went up. Normally shareholders do not lose money when the stock price goes up.
It's weird that they had to report it like this because they are technically derivatives, and if they had just given out actual shares (as with stock-based comp), they wouldn't. Yes, giving people stock is a real cost, but what happens to the price afterwards is not, and when the price goes up shareholders are definitely making money, not losing it.
Either way, they did report it. It is up to you to decide how you interpret that information.
Their bet is that they will acquire other companies to become valuable ( their ceo said they are competiton to every other industry)
It does not make sense but it is what they say
In some ways it is crazy, but at the same time, anyone that has seen typical efforts of "digital transformation" in aging Fortune 500 companies is going to be skeptical of said company's abilities to become relatively efficient. I always considered that new entrants that don't think like they do would eventually win once the value of putting tech in the middle of decision making, getting rid of large divisions worth of cruft was worth enough. Trying to do the very same thing by competing with everyone is similar once you squint. You get extra money from your parent company to throw around during growth or massive reorg, but still have the problem of iffy incentives we currently see in the modern tech holding company, where what is good for the company, for consumers, and for the random VP have nothing to do with each other.
you can't acquire your way to value, because the seller will want their piece when they sell, and eat up that value as their premium for selling.
The only way to get value from acquiring other companies is to buy them when they're distressed.
That's why the playbook is to force the companies into distress.
I'd like to see the list then.
I don't fully agree with you that it doesn't make sense for Anthropic to claim it is in competition with every other industry.
Take a look at how OpenAI has shaken up higher level math.
Frontier AI labs have a Swiss army knife of a tool that can do as well or better than every other company or collective in every industry.
Say you have a magic hotpot that can deliver any dish you want and each dish it delivers can be as good or better than any restaurant in your area of business. You basically cannot be beaten in the food business. Your competition can only cry about their cooking being more artisanal or whatever but at the end of the day convenience and cost will triumph.
Now say you have a magic pen for art, a magic jukebox, a magic text generator for authorship, and so on. You get the idea why I don't think Anthropic is wrong.
I guess AI will be very competitive in all research fields, but your example is actually interesting: they do not have a magic hotpot, right? So are they competing in the food industry? Or how quickly will they be able to compete with Amazon in logistics?
So I think they can compete in many industries (and I don't like it), but not all of them for sure. They can only compete in all industries with infinite money, and I don't think it would be good for most people.
Anthropic (and OpenAI) has an amazing opportunity to drive the value of multiple other industries to zero, yes. I'm just not quite sure how it benefits them.
Art, music, writing... those are all cheap. They've also already got plenty of good models out there, and people will judge you for using them.
The places where American society needs costs cut or needs an increase in labor - healthcare immediately comes to mind - are places where you'd be asking Anthropic (or any other AI player) to take on massive amounts of liability.
If you screw up someone's drawing, you regenerate the drawing and charge the hack/fraud for the privilege. If you screw up a blood bank order, you pay the survivors a few million dollars. Since you've created a system that can scale blood bank order up-screwing, you'll be doing that a lot. There's not profit in paying out medical torts.
Given that SV-based companies work to dismiss liability even more than most American companies do (which is saying a lot), I don't see that working well for them.
Have it build an airplane.
Why aren't they aiming for a reasonably sized valuation then?
Because if they priced the company more reasonably many later stage investors would stand to lose money. So they’ve backed themselves into a corner of having to go public at a $1T plus valuation just so some later stage investors don’t realize a big loss.
Their last investment round in May had them valued at 965B. And the company has grown a lot since then, those investors obviously expect a profit
From where? There may be some profit but I don’t think it’s gonna be anywhere near what they are expecting and the same goes for OpenAI.
They are probably going to need all that capacity.
Maybe they can use it, but if there is an overshoot of investment in compute, future competitors may get it for firesale prices while Anthropic pay more than full price. That is, they can get dragged down by those commitments not because their own growth is lacking, but because the market gets weaker. They may even contribute to it, the more competitors they kill, the less demand there is for compute.
To me, "a probably $200-300B company" sounds absurd. You're describing one of the possible outcomes of a massive range that goes from $0, where they get over their skis on financial commitments to well beyond $2T, where their 10x YoY growth continues for a few more years. A $200-300B outcome is possible, but certainly not probable. (And in fact, for us HN commenters, the most probable outcome should be assumed to be roughly what the market will bear.)
>The issue is this is a probably $200-300B company pretending it’s a $2T company.
What do you expect for FY2027 revenue to confidently conclude this? not saying you're wrong, but considering the dramatic changes in their revenue you seem strangely very certain that you know where it will land.
So is SpaceX but markets don't care.
Also, anthropic has no real moat, LLMs are a commodity.
Unless they are able to create God. There is no possible road to profitability.
> The issue is this is a probably $200-300B company pretending it’s a $2T company.
They probably asked Claude. Sounds like a normal LLM hallucination.
Should have followed it up with a second prompt to get the real answer ... "You're absolutely right. valuation is only $20b"
From a financial analyst on Bluesky: https://bsky.app/profile/peark.es/post/3mzdkyxzzns2w
> Okay so Q3 ARR at $67.8bn versus $46.2bn for Q2, +47% QoQ. Turns out tokenmaxxing's end over the summer slowed revenue growth versus an absolutely torrid 100%+ growth rate in Q1/Q2 but 46% sequential is still eye-popping
> Operating margins (GAAP) went from 8% to -3$, with a -28.5% QoQ incremental operating margin. $SPCX compute coming online is the reason.
> Bottom line after seeing this Q3 update: the basic takeaways from the original S-1 don't change, but fears that they were excluding Q3 because it was a disaster were definitely wrong. Do with this information what you will.
It's also worth noting that Q3 only has one week of Opus 5.5 (9/22 - 9/30) which may have changed the economics for Q4.
> fears that they were excluding Q3 because it was a disaster were definitely wrong
That's a very generous reading. Anthropic's entire growth narrative is the rapid growth over the previous 12 months, from near-enough zero to $100bn+ this year. Now we see that actually growth is levelling off and that their revenue was ~$5bn/month over the last quarter, which is barely over their compute commitments, and nothing close to the $100bn+ leaked.
I guess if people thought Q3 would show negative growth rate, sure, it isn't a disaster, but their growth rate slowing so much is surely very bad and matches all the concerns commenters had.
And this is Anthropic! Anthropic are the financial success compared to OpenAI. OpenAI are doing much worse, so this does not bode well for OpenAI either.
People will point out that +46% is still good, but this is like measuring the escape velocity of a projectile that's still at 1000 feet. It's not that the projectile isn't going very fast or won't get very high, it's that if you want to have enough momentum to break out of Earth's gravitation pull you can hardly afford any deceleration.
Businesses aren't projectiles
Their forward pricing prices in all expected growth... so if they don't have accelerating momentum that reduces expected price.
You're now talking about the expected trajectory.
Really, you're talking about expectations ("expected") of the expected ("price") trajectory (actual business performance).
If the above comment was intending to say "chart go up steeper make price higher, chart go up less steep make price lower", then no dispute there. But that carries virtually zero predictive information about whether the business will achieve "escape velocity" (whatever that means).
You absolutely cannot predict Anthropic's fate based on their growth rate slowing from Q2 to Q3. Obviously.
You can predict the resting place of a projectile with information like this. You cannot do so with a business, especially one as dynamic as Anthropic in an environment as dynamic as today's.
The point being made here I think is that the same can be expressed in natural english language that does not pretend to be a physics class for some reason.
That reason possibly being to grant legitimacy to an otherwise mostly illegitimate and equally mostly fictional branch of the economy (that whole MBA cluster).
Mathematically they can work very similarly though because you have a revenue growth rate, and margin. The question is can that revenue/margin reach break even before you run out of runway (the expected sum of all rounds of investing) .
The main difference I would concede is that gravity, mass, forces, awind drag etc are much easier to model.
Sure they can work very similarly. But they can also not behave very similarly.
Businesses can have stalled growth, degrading margins, etc. etc. and then simply... reverse it. They can raise more money. Their competitors can get regulated out of existence. They can do a literal infinite number of things that cannot be factored into a model like "a projectile."
That's why the projectile analogy is stupid as fuck. It suggests you can presuppose the conclusion by presupposing the conclusion.
It's a model that helps you make present tense valuations of potential futures.
No one said it literally is a projectile. It's an analogy and model. The similarities are in the present tense maths. It's presumed and beyond having to be stated that businesses exist in a complex set of variables all of which could change and might not be as predictable as the valuation maths.
You're not wrong, you're just not adding anything to the conversation that any but the most uneducated do not already know and acknowledge.
And the analogy is literally not useful in any way whatsoever.
It implies that you can know Anthropic's outcome from this slice of data. That is literally not true in any way at all.
Like it's not even remotely or analogously or roughly true.
Anthropic behaves like a high-growth business with a quarter of slowing growth. It does not behave like a baseball. Any conclusions you draw from analyzing it as a baseball will be de facto faulty because the two systems are nothing like one another. The most basic assumption of projectile analysis (that energy is not added in-flight) is the opposite of the most basic assumption of a business (that they are actively profit-seeking).
The analogy adds nothing except ambiguity at best and active falsehoods at worst.
>It implies that you can know Anthropic's outcome from this slice of data.
No it doesn't. All I said was that given how little we have to go on and how early in the process we are, a minor 2nd or 3rd derivative slowdown isn't a positive signal. All indications are they they need additional orders of magnitude in revenue fairly soon.
I also never said that it wasn't self-propelled. I was starting to write about waiting to see if another stage of propulsion comes online but I deleted it.
Okay, well if you weren't suggesting you can predict the outcome based on this slice of data, then "projectile" is a really stupid analogy. The defining characteristic of a projectile is that you can do this.
Analogies however are: they often fly over people's heads.
The analogy fails at precisely the thing it's attempting to do.
It's saying "like a projectile, you can predict the outcome based on just a sliver of data"
That is literally not true.
Do you respond to every metaphor this way?
Bad, misleading, and uninteresting ones, yeah pretty much.
Businesses also aren't financial statements. These are abstractions we use to understand and communicate about the world.
Correct. Which is why making a physical claim about a 2nd order derivation of a super dynamic system is goofy and misleading.
But they do have S curves. If you're on the shoulder of the S curve while your spending commitments are accelerating, that won't end well.
Presupposing where they are on "the" S curve is itself presupposing the conclusion, duh.
Most businesses are stacks of S curves (except when they're not).
AGI is supposedly the literal be all and end all. It's a bit embarrassing for the singularity to have an S curve.
Someone needs to tell that to the AI startups.
I think that's an unnecessarily pessimistic take. The way I look at it the industry did not wake up to how big this was until Q4 and so it makes sense that you would see astronomical growth in q1 and Q2 as everybody woke up and started adopting. Seeing the growth Taper down to a 4.5x annualized rate which to your point will only continue to taper but it's still mind-boggling at this scale ... you just don't have anything else like this in the economy. This is like hyperscaler Revenue scale achieved in 15 years being achieved five times as quickly. At this rate Anthropic will exceed AWS Revenue within a year
>At this rate Anthropic will exceed AWS Revenue within a year
And in 10 years it will be worth the entire Earth's economy.
https://imgs.xkcd.com/comics/extrapolating.png
It is a lot easier to switch model providers than it is to switch cloud providers. And the open weight models from Chinese Labs make Claude a lot more of a commodity than any service provided by AWS.
> At this rate Anthropic will exceed AWS Revenue within a year
Well that’s the point. They’re not continuing at that rate of growth.
10x year over year couldn't continue for long--they were what...3.5 years from exceeding US GDP?
47% quarterly growth is still 4.5x annual growth, which would put them at $300+ billion revenue in a year. This is still unprecedented growth for a company of this size.
The negative perspective depends on assuming that the 47% quarterly growth is going to continue dropping soon. And it is going to drop, since it's only a bit more sustainable than last year's growth. The question is how many hundreds of billions they're making a year when it levels off.
> Now we see that actually growth is levelling off ...
I really don't understand how could it be otherwise? They're planning to IP at a valuation representing 5% of the US GDP and, unlike NVidia, they are neither selling shovels in a goldrush nor a business with decades-old proven track of records.
Are people really going to buy ANT thinking it'll 10x to 50% of the US GDP?
Aren't there better stocks to YOLO on if you think you could maybe 10x? And if you want something safe instead of YOLO'ing, aren't there plenty of companies much safer and much more boring to invest in than something IPO'ing at 5% of the US GDP?
Isn't the entire Anthropic narrative based on infinite growth? How is infinite growth sustainable?
> Aren't there better stocks to YOLO on if you think you could maybe 10x?
This is why any [consumer] investment in ANT is bad. If ANT goes 10x then congrats, we [the citizens] are in a crisis. If ANT goes 1/10x then congrats, holders of ANT are in a crisis.
Very generous indeed. Here is the updated chart. The downward movement should not be overlooked.
https://x.com/billylo/status/2117276235852345425?a=4-pjgf1Ox...
A single downward point for one quarter is not a "trend."
I did not suggest a downward trend. (It's a trend graph)
If we assume LLMs will continue to advance at a pretty rapid pace (relative to the current rate of advancement; so, same or faster), all these numbers look good to me.
Some people just start out with wanting LLMs/AI/Anthropic to fail or be scams and consistently try to fit any new evidence pegs into those unchanging hypothesis-shaped holes. This group ideologically despised LLMs from the start and needs them to be fake.
Lots of people making the financial case against Anthropic are not anti-LLM. I think the technology is great, interesting, powerful, innovative etc etc etc my issue is with the garbage business that Anthropic and OpenAI are building and the nonsensical ideas people have about business in general. Technology is exciting, business is not, business is boring. Money come in, money go out. Selling technology, no matter how exciting it is, does not change the nature of business.
And yet one can argue that the "garbage businesses" you label are the ones doing their best at matching market demand. See how well things are going for other entrants, like Google, Microsoft or Meta. Putting Copilot in every app makes a whole lot less sense than Cowork, Codex and the like. They just are doing worse things while attached to traditional money fountains, like ads in social media and search.
It doesn't matter how much they "continue to advance at a pretty rapid pace", if the economy has no actual demand for the new capabilities. Even if we accept the premise that LLMs can produce good code at scale now and automate most software development work, then what? As *an industry*, the bottleneck was that we needed people to write code, and that bottleneck has been removed since ~Opus 4.6. Where is the industry now bottlenecked which requires some significant quantum leap in model capability?
Opus 4.6's coding ability is nowhere near as good as Opus 5.5's. Some model next year will almost certainly be significantly better at coding (and planning, and designing, and ideating) than Opus 5.5.
You're narrowly focused on writing code. What about when they can run entire firms themselves, better than the average human CEO could? What about when we have models way, way, way better than Opus 4.6 and 5.5?
The point is we will eventually very likely have AI systems that can do everything. Everything. Anything and everything a human can do with - or even without - a computer, and then better than anything or everything a human can do, including in terms of creativity and innovation and novelty and insight. Obviously I don't know when that might possibly be, and it's not happening next year or something, but you can't just pretend like we hit the cap on the technology.
> What about when they can run entire firms themselves
That would mean the whole idea of free markets will be rendered obsolete, instantly destroying the idea that investing in Anthropic is sensible.
Don’t bother - these knuclkleheads don’t understand nuance and despite much of their devices being manufactured in China - they struggle to accept Chinese offerings can take significant market share.
Not if Anthropic consistently stays 6+ months ahead of all labs besides OpenAI.
>If we assume LLMs will continue to advance at a pretty rapid pace
Why assume that?
So much money have been given to that handful of companies and to that field, that it's pretty much guaranteed to plateau (it's not some small but growing market with lots of room to grow. It's already taking as investment a huge chunk of the economy, and the returns don't seem to justify it).
Most of it is a pre-IPO hype and a circular financing loop between Nvidia, Oracle, Amazon, and the AI players anyway, so the real returns are even worse.
You can be bullish on LLMs profoundly transforming society while being bearish on the economy "digesting" this new technology fast enough to justify the incredible capital spending necessary to power it.
I am actually somewhat bearish on adoption/propagation. I would be expecting much larger numbers than these if I were bullish on it.
No there are just numerate people pointing out ANT/OAI need to make X $$$ to make Y $$$ spend make sense. The financials are separate from model capabilities, they can cure cancer and invent big dick pills and still go out of business if their balance sheet is fucked.
Yes, but their balance sheet seems fine to me.
I don't really follow how LLMs advancing leads to OpenAI or Anthropic being successful businesses? Those seem like orthogonal outcomes.
They're certainly correlated, even if one might quibble about the coefficient.
If you have no theory of how to value a company any number looks good.
>If we assume LLMs will continue to advance at a pretty rapid pace (relative to the current rate of advancement; so, same or faster
https://xkcd.com/605/
Yes.
Disagree on the tokenmaxxing end. People are still doing it (and I'd argue more people are doing it due to wider adoption). Its just that most low hanging coding adoption has reached saturation point very quickly (which is genuinely impressive).
Now, they need to unearth new use cases and saturate them to keep the momentum going. Its unclear if they have found these opportunities and can turn them into sustained growth.
No they are not. Big tech companies started capping AI usages already and those are their major clients.
Correct, and most every company is optimizing around model routing, with many including much cheaper open weight models.
Within large tech companies the era of "spend the most money on the smartest model running on its maximum effort, always" is dead.
Yeah I think for the last 3 months or so, the only ones that are still token maxing are the ones on subscriptions.
Doesnt matter at this stage of growth what financial analysts say because they havent seen anything similar in their lives :) Large corps are going to really be the deciding factor, in what ANTs next years revenue is going to look like. And right now inside large corps there are literally hundreds to thousands of different experiments running and no one knows what is going to click.
The takeaway is there is no moat and people will switch to competitor model in a heartbeat.
That's not the takeaway of this analyst or anyone I have heard.
This is specifically about the end of tokenmaxxing and how much worse the growth rate would look after the Q3 numbers came out.
(I do think moats are shallow, and I personally switch away from Anthropic every few months for a while, but I've been coming back. All software moats have been made more shallow now, affecting the entire investment landscape.)
Models are not a moat for sure.
Anthropic is working hard to create a vertically integrated stack from model to harness to productivity apps. That may or may not work, but it’s clear they see models as a value add for integration and not a product that can be a moat in themselves.
harness is not a moat. CC is great, I mix it with codex with my own vibe coded messaging and hierarchy layer. Lately there are more CC because of Opus5.5 but Codex caught back up with Sol-6.1;
Codex also wins by consuming a lot less memory.
The product doesn't have a moat, but anthropic's branding and enterprise focus seems to be quite valuable.
I disagree. Our devs use Claude for coding. When OpenAI had Astra, no one switched even though the company allowed OpenAI models.
We have Copilot and can select between various OpenAI and Anthropic models, and when I accidentally select the “wrong” one it takes me awhile to notice.
I use Claude Opus at work and DeepSeek v4.1 Flash on my personal projects. For the kind of work I'm doing, they're basically interchangeable and DeepSeek costs 3% what Claude does.
https://openrouter.ai/compare/anthropic/claude-opus-5.5/deep...
How does it change if you’re building via subscription?
what if the alternative is much cheaper?
Arguably, it was cheaper.
By the way, we get way more value from our Anthropic subscriptions than any Chinese subscription.
How do you define value?
Opus 5.5 is great, but you are kidding yourself if Deepseek 4.1 flash isn't as good as good for 99% of coding tasks. Anthropic (and OpenAI) are between a hammer and an anvil in terms of good enough local models being able to run performantly on current and next gen Apple hardware.
Your dev’s are not representative of everyone bozo.
Your posts are irrelevant.
Does this person not own a calendar? Q3 numbers still polluted by expensive tokenmaxxing, which wind down late Q2 / early Q3 i.e. mid year budget review which will take rest of year to unwind. Q3 numbers = more mirage of tokenmaxxing tail and velocity = lagging indcator of tokenmaxxing end. Q3 also when fable tier coding commoditized and everyone doing a deepseek on cheap inference. The latter is what will kill ANT hardest, they signed up for 100s of billion of none cancellable compute contracts last year based on projected demand x expensive token cost, but demand / growth is less than token max timeline while inference cost getting cheaper. They have fixed, exorbitant costs, and operating margins negative for Q3, and Q3 just beginning of full price war. TLDR they will likely be vastly overpaying for compute. If turns out jevons isn't people using more expensive tokens on more expensive compute, but more cheap tokens on less compute, then ANT's entire balance sheet breaks.
>Q3 numbers still polluted by expensive tokenmaxxing, which wind down late Q2 / early Q3 i.e. mid year budget review which will take rest of year to unwind.
Evidence? I was told by a lot of people that token use was going to start to slow down 5 months ago because of "the end of tokenmaxing."
openrouter still showing weekly token usage higher than ever: https://openrouter.ai/rankings
People PREDICTED token maxx was a fad/trend Q1 while it was underway, but companies asking employees to cut back was late summer, i.e. Q2/Q3 transition. And the argument isn't end of token maxxing its end of EXPENSIVE tokenmaxxing, i.e. tokens pushed at $$$ and volume that can make ANT numbers square. They're not precisely because people can now maxx on cheap commoditized tokens. The market for premium tokens (at least in SWE) got kneecapped, there's only so much demand for smart/fast.
>People PREDICTED token maxx was a fad/trend Q1 while it was underway,
No, a lot of people like Gary Marcus said it had specifically already ended in May/June. Why is this time different?
>Tokenmaxxing is over. https://finance.yahoo.com/sectors/technology/articles/tokenm...
>Tokenmaxxing is dead, long live valuemaxxing https://www.ibm.com/think/insights/tokenmaxxing-dead-long-li...
>Token Maxxing, April 2026-May 2026, RIP https://x.com/GaryMarcus/status/2057610132742156603?lang=en
Yes look at the dates... end of Q2, i.e. start of Q3, things take time to unwind. Marcus tweet MS cancelled contract, but doesn't say when terms end, but they almost always unwind over following monthS not ASAP. Real world works on a lag / budget cycles not how fast tweets / articles get published.
> Okay so Q3 ARR at $67.8bn versus $46.2bn for Q2, +47% QoQ. Turns out tokenmaxxing's end over the summer slowed revenue growth versus an absolutely torrid 100%+ growth rate in Q1/Q2 but 46% sequential is still eye-popping
There is an unbelievably enormous difference in 100% vs 47% -> and when you look at the 2nd derivative this is REALLY REALLY REALLY bad.
I mean, the growth is astounding and incredible. It does not bode particularly well for a ~$2T valuation.
This is the correct analysis here. 47% QOQ growth IS objectively amazing. But it's not enough for a $2T valuation.
> It's also worth noting that Q3 only has one week of Opus 5.5 (9/22 - 9/30) which may have changed the economics for Q4."
An average random dev here: I switched to Opus 5.5 and forgot about all other (paid) models.
From a loud AI booster on Bluesky. Man, I am relieved now, the science is settled.
Booster is an exaggeration.
We desperately need to ground these discussions in actual numbers.
This should help build some intuition and establish a baseline for what the $2T valuation means.
If you buy at $2T Anthropic needs to be worth about $3T by 2030 for a normal market return. That's ~25x P/E and ~$120B/year of profit.
That could be something like $400B of revenue at 30% net margins, $300B at 40%, or $600B at 20%.
It would take 1.7x revenue growth a year to get from today’s $68B run-rate to $400B by 2030. Anthropic's current growth rate is ~4.7x a year. Anthropic doesn't need to maintain that, they basically just need to prevent it from absolutely collapsing and hope margins hold up against open weight models.
You could look at this and say "yeah that's plausible" or "I highly doubt they achieve that" but it's hyperbolic to say its actually impossible they can meet this valuation, definitive proof of bubble, cant be worth this unless they can create God, etc.
> [Anthropic] basically just need to prevent it from absolutely collapsing and hope margins hold up against open weight models.
There are so many variables involved that are outside their control. If GPU/RAM demand dries up (or supply increases), for example, local models are suddenly a threat in addition to hosted open weight models.
The quality there is already good enough. I've been 100% local for personal projects since Qwen3.8 Flash dropped. It's not as nice an experience as Claude I use for work, but it gets the job done and I fully control my data and can do whatever I want with it.
You just made me wonder if part of the buy-up of RAM by OpenAI and other spend on GPUs and datacenters is literally trying to weaken the value proposition of open weight models.
Not just against open weight models and not just on hardware I think. Also on token subsidies etc. OpenAI and Anthropic both seem to be aiming to outspend everyone else, so that they can be the last ones standing. It is a risky move, especially since everyone else includes Google and Microsoft, players that have both existing surplus cashflow, solid warchests, and excellent distribution channels. But all outcomes are open still...
They are absolutely trying to deny their competitors necessary hardware, or at least drive up their costs substantially. And users running local models are, well, not not competitors ...
Yes people were saying that since a year ago when Altman's deals with the memory companies were first announced.
Interesting. Their revenue has "only" gone up 47% in Q3, compared to doubling revenue in most previous quarters. Meanwhile their "compute and infrastructure" cost has increased by 77% compared to Q2
Maybe they tried too hard to get good Q2 numbers, making revenue happen earlier and expenses later, and that now makes Q3 look worse? Either way, this does not look great
They clearly mention the cost increase in the amendment - it's because of the spot pricing of gpus (likely SpaceX). Their demand exceeded so much they had to quickly find some gpus at _very_ expensive prices
And they say that this will likely continue to grow until 2027 (when the long term commitment compute comes online, which comes with opposite problem .. will there still be enough demand)
It still doesn't look anything like a death knell. There is a point where you approach market saturation and the growth curve follows a sigmoid. Have you ever tried folding a piece of paper ten times?
I agree. Their financials are fine. They look like a company that has solid foundations, with room to grow but close enough to profitable that they can survive the end of the AI boom
But are they still a 1.5T USD company if their growth is already slowing this much?
We’ve known for a long time they goosed Q2, and we even know how.
Apparently you’ve missed all of Ed Zitron’s reporting?
When was the last time Ed Zitron got a prediction/claim correct?
Irrelevant, his numbers and analysis is where the value is and those are hard to dispute.
What are his numbers and analysis?
So he can get all of his conclusions and predictions wrong based on his numbers and analysis, and then you think his value is in his numbers and analysis?
Required reading if you're citing Zitron: https://danluu.com/zitron/
Funny how no one ever disputes his facts, only his predictions. He can easily be wrong on the predictions, while still being correct about the facts, i.e. their financials are shit and completely unsustainable. I don't care at all about his predictions, and am very interested in his fact-finding.
What is unsustainable about Anthropic's numbers?
Can we all please stop acting like revenue is somehow an indicator of a successful business?
If you are trying to indicate that revenue is somehow not a key indicator of a successful business, or that somehow revenue growth is not the primary indicator used to derive Anthropic's eye-popping valuation for IPO, then I think you have lost the plot.
Revenue alone is not enough as we saw in the dot com crash. Scalable, profitable growth is required.
That is true, but not even really connected at all to the comment I was replying to. Revenue is indeed an indicator of a successful business.
Additionally, the Anthropic S-1 does indeed show scalable profitable growth. This is perhaps the most shocking thing in the S-1, to me.
No, not when that business spends more than it makes to bring in that revenue.
A better way to phrase your thinking might be that revenue is an indicator of demand.
The issue here is that they've not really tested what happens to demand when the prices moves to where it needs to be to cover the cost of keeping the lights on.
See here: https://news.ycombinator.com/item?id=50315574
>People should really look at cash flow statements instead of just earnings. Anthropic published one and their cash flow was positive for the first half of this year.
The dot-com era had a lot of companies with high valuations and zero revenue. That was the main problem.
If you go just by operating income, they showed they can be profitable in Q2, and the operating loss in Q3 is fairly small (in relative terms. For every 32 dollars in revenue they lost one dollar). They haven't shown wild profits, but nobody expects them to. Any money they make has more value when it's reinvested into growing their market share and growing the market, making next quarter's numbers look better. And imho they have solid enough financials that they can make the bleeding stop if the money faucet turns off
All of that isn't very newsworthy though. We knew all that from the Q2 numbers. What is interesting is that reinvesting all their profits to make the number go up in the next quarter didn't work as well this time as it did all the other times
I'm unconvinced they have a moat that will allow them to ever show wild profits though. They are constantly 6 months ahead of open weight models. Their potential customers are perfectly able to wait them out. They'd need a government enforced cartel against open weights at this point for me to not view the IPO as a plain and simple cash out.
Some of my nephews ran a hugely successful pokemon card selling business (for ten year olds). I later found out that their mom was buying all their card packs for them to open and sell.
They stopped when the free packs stopped coming and they needed to finance their own.
Revenue is the one number that indicates how interested customers are in your product. What would you have us measure, then?
There is no company on earth that can run on customers being interested alone, you also need to offer a product they can afford at a prices that generates profit.
Right now, I could start a business selling gold bars for $1 a piece. There would be more people interested in that than Anthropic.
Yes, revenue shows people's interest and willingness to pay for the product. That's exactly what it is.
ANT also has positive gross margins, so they aren't selling gold bars for $1.
I don't understand why people on HN treat ECON101 like some fringe theory every time this is discussed. This is basic stuff.
Gross margins just means that they make more in revenue than their operational costs to offer the product (i.e., electricity, salaries, etc).
They have yet to release a model that has made more in revenue than it cost to develop...and each new generation costs more to develop than the preceding generation. Costs increases outpace revenue increases with every generation, across the industry, and LLM-based cash flows are highly negative even at Google and Microsoft.
Customer use of LLMs is decreasing after quarters of growth. Execs were all in on pretending LLMs were the future until the WSJ/Stanford report revealed what everyone outside of tech already knew: LLM use has not correlated with any measurable productivity growth in any non-programming industry. (And arguably, the same is true in tech as well: LOC has increased but quality has dropped, so the net change to productivity is de minimis.)
I like keeping an eye on OpenRouters view of the market: https://openrouter.ai/rankings#market-share
Since the start of the year the market has shifted from 50/50 split (closed v open weights) to 30/70 (70% of usage they see through their platform - is now open weight)
DeepSeek V4.1 Flash 1017 sees 5x the weekly token usage through OpenRouter than the closest closed weight model (GPT-6 Luna)
Even though OpenRouter is only one limited view of the market - those trends should be worrying to potential investors.
I'm not saying that OpenRouter isn't worth looking at but it, obviously, shows a much smaller slice of LLM market and it's going to, obviously, trend towards open models since those are the only model where there is any competition for inference pricing. No one is undercutting Luna because they can't really.
OpenRouter could so 90/10 split and it still wouldn't mean open models are beating (in usage) closed models.
Looking at the numbers there though, both Anthropic and Deepseek increased in terms of absolute usage numbers since the start of the year. And thats setting aside the high potential for sampling bias.
This view will ALWAYS overweight open models. The closed ones are being used in their own harnesses via direct API connections that OpenRouter will never have visibility to.
How do you explain the trend changing since start of the year if that were the case? Closed source models were leading open weight ones before Jan 26
Let's say I'm starting a shitty cash-grab SaaS that needs some kind of "Ask Jenny" feature. Is the standard practice to sign up directly with a Claude/OpenAI API account?
We need token monitor apps to reveal the less biased stats.
The sentiment they are trying extraordinarily hard to create is that in the long run Anthropic will be the only one with any AI worth anything.
That makes any loss or risk reasonable.
It's great propaganda, so great that they have themselves started believing it with certainty. This is one of the self owns that brought down the USSR :)
Of course they could be correct too, time will tell.
Beyond it being propaganda, they're trying to work the regulatory capture angle to pressure lawmakers to pass safety regulations that would functionally ensure a market in which there are under a dozen players legally capable of competing.
It's more accurate to say Anthropic intends to scam retail investors by claiming they will have the only A.I. worth anything.
Then if all goes well next year OpenAI will scam retail investors by claiming they will have the only A.I. worth anything.
And of course, obviously untrue to anybody paying any attention to the rapid progress in open weight models. The IPO will be a cash out for insiders and leave the typical institutional investors with the bag.
It could be untrue, but it's not obviously untrue.
The open weight models right now are nowhere close to ..say.. Fable 5.1
They could catch up if Anthropic reaches some sort of a plateau and they manage to keep improving, but there's no reason to be certain about that happening.
dead wrong, kimi k3.1 is good
"Just don't look at those cheap af Chinese models pls"
Genuinely think google + Meta might be the only companies to make a profitable AI product at this rate.
Google currently appears (by their latest numbers) to be sacrificing the only substantially profitable portion of their business to attempt to create that reality.
You're assuming costs remain high.
Pretty much every technology in history is available at several orders of magnitude lower cost than the first few years of it being invented.
AKA eventually everything runs on a potato.
> AKA eventually everything runs on a potato.
If we're talking web servers, sure. Video games? For the most part. Numerical high performance computing? ehhhhhhhhh i'm not sure history bears that out
> Numerical high performance computing?
Any numerical high performance computing from 1995 will run on a 100$ android phone.
Really? Even with this very generous 30 year gap ... you think you can run non-toy climate models on a $100 phone? Correlated quantum chemistry for a polyatomic system? Even just fully ray-traced rendering?
How much of a potato? You think we will eventually be able to run inference on a rasp pi?
> You think we will eventually be able to run inference on a rasp pi?
Not today's raspberry pi, but the equivalent of it absolutely. I don't work in it, but have a number of friends who are involved in LLM development - the "breakthrough" optimizations are obvious work in hindsight. We are still in the very early innings of optimizing both hardware and software techniques. The recent market imbalance has blurred the idea that hardware simultaneously gets both cheaper and much better, but no one serious expects that to stop on a longer timeline.
Going beyond the "my watch has more computation power than NASA at the moon landing" - speech-to-text used to be unreliable tech that needed datacenters backing it as recently as the early iphone, and now runs in near-realtime on smartphones that cost about the same. A ~$50 processor running in my closet does realtime person/vehicle/package detection on my front door camera reliably.
Interesting point. I hope you are right and that time comes soon because I'm very much regretting not building a new computer in the last 10 years.
Whatever you can buy for inflation-adjusted-45$ in ~2035, almost surely.
Possibly. 1.58-bit models already run on CPUs https://github.com/microsoft/BitNet They're not good exactly but they are surprisingly better than you might think.
https://huggingface.co/microsoft/bitnet-b1.58-2B-4T#evaluati...
~1 bit models are pretty much decision trees in disguise! :)
Which LLM use cases can be solved with compact decision trees is an awesome open problem.
Meta? They're not even in the game. They don't provide inference to anyone. I think NVIDIA or any other hardware players that want to step up and take the bags of money awaiting them are the only actual winners. They also win when we inevitably move to local AI for many use-cases.
Meta provides inference to itself, and the 3.58 billion DAUs across its services.
https://searchengineland.com/guide/facebook-users
This is them working within their existing market. This isn't the new market that the valuations for companies like Anthropic are operating in.
Meta are using AI internally to target and sell ads more effectively.
Consider: if you had built a hyperintelligent stock-trading AI, would you:
- Charge $0.70/M tokens for other people to ask the bot what stocks to buy.
- Get rich following the bot's trade advice and never tell anyone about your bot.
> Get rich following the bot's trade advice and never tell anyone about your bot.
The fun thing about open markets is the market will punish you for this with second order effects and it's in your best interest to get some whales on the bandwagon.
This is the existing ad market with the existing TAM. Not the AI market.
"It's not about how much you earn. It's about what you're worth. And who's worth the most? Companies that lose money." ~ Russ Hanneman
Leaving aside that Meta are selling compute. The reason I pick on Google + Meta is they are able to make their own customers + charge their own prices. Anthropic + OpenAI are both reliant on people coming up with Products powered by their AI models valuable enough to create several AWS worths of revenue and also not also not immediately moving it to open models.
the only revenue stream google has is selling user data & ads. always has been, always will be
It’ll only be profitable because they’ll enshittify it with ads.
They also have the second mover advantage and the financial stability to put it to good use.
Hate all you want on those two (I certainly do) but they are incredible at monetization
Omitted/minimized: No mention of its models creating security issues in new code and the mass of vulnerabilities AI agents are creating in its wake. If you run Fable on Opus code, how many security issues do you find?
See: https://news.ycombinator.com/item?id=50157976
AI Coding Agent Security Benchmark https://www.endorlabs.com/research/ai-code-security-benchmar...
Natural Emergent Misalignment from Reward Hacking in Production RL (https://www-cdn.anthropic.com/daad4360a8bdc707f8b22e3e745796...)
Surprised these financials are not too dangerous to release.
They were, but then Anthropic got so much flack in the business press for omitting them publicly (while sharing them privately with potential subscribers) that the bad PR was as bad as the financials themselves.
Yea, my guess is they decided they would rather have the actual Q3 numbers (and any reactions to them) rather than anxiety about unreleased Q3 numbers priced into IPO performance.
Which was probably the right move. The S-1 didn't seem to really change the overarching economic and financial narratives about Anthropic (it seems like whatever people already felt before the S-1, they felt it even more after looking at it).
So I don't think there's going to be any problem with boosters metabolizing the Q3 numbers.
(Idk, maybe this was their plan from the jump.)
My guess is they already heard from big investors they weren't going to participate unless they had the numbers.
They asked Claude to evaluate the risk of annihilate the human race by releasing these reports.
It says "You're absolutely right!".
So they did release the reports.
Direct link to the prospectus: https://www.sec.gov/Archives/edgar/data/2130357/000162828026...
Consolidated financials are on page 23, quarterly numbers on page 160
I am not sure how the IPO is going to go.
Tokens themselves are already a bit of a commodity, and the margins are gonna be fierce on these as there is not a whole lot of lock in.
Would Anthropic and OpenAI position themselves using their extra tools, their ai agents, presentations, docs. Are they nice; yes. However, that doesn't make a 2T company. It reminds me a bit of Dropbox, a truly great product, however, they're differentiation is simply not there. I think the momentum is there for the IPO to go decent, but it feels like the company is on such unstable ground that the fallout could be catastrophic.
I think the unstable ground comment applies to all purely AI companies.
Google's strategy right now seems great.
Agree. Was starting to look seriously into Google's AI stuff last night.
Owning the best search index and being able to retrieve content immediately straight out of their index seems like such a massive competitive advantage.
FWIW, from my prompt last night: https://chatgpt.com/share/6ae8d441-8c44-83e9-980c-b14a2ed846...
The market is completely divorced from fundamentals at this point. Just look at Musk's companies. Neither Space X nor Tesla have the fundamentals to justify even a fraction of their valuations, yet here we are. Anthropic wants in on the same bags.
If OpenRouter is a proxy, then their growth will pick up again in Q4: https://openrouter.ai/anthropic
In Q3, I think they struggled with the government ban on Mythos/Fable, Opus 5 not being so good, OpenAI models being more token efficient, and they decided to dedicate a higher percentage of compute to training.
The increase in token volume corresponds with a price cut. The S-1 revision already reflects what's going on financially. Anthropic is buying volume with margin. There's nothing wrong with that but we shouldn't expect profit to show back up as Anthropic's costs are increasing and have a fixed minimum. Given that OpenRouter users can switch models with the click of a button, what this demonstrates is that OpenRouter users are extremely price sensitive.
Yeah, I was thinking the same thing. Fable 5.5 is coming out soon as well.
Openrouter is a terrible proxy for frontier, token leaderboards are dominated by fast models like Deepseek 4.1 or Luna 6.
Yes, but the trend is pretty clear based on the link I posted.
Hahaha cringe.
It’s not possible at all that they had to cut prices to hold market share as folks switched some of their usage towards open models.
Utterly cringe takes.
Net loss of $23b in the third quarter is pretty bad right? Was that also due to the fair value of convertible notes?
About $22B of it, yes.
For those looking for the actual S1: https://www.sec.gov/Archives/edgar/data/2130357/000162828026...
and add'l discussion: https://news.ycombinator.com/item?id=50338510
and for those who want it converted to a legible PDF: https://enopdf.com/tools/edgar-to-pdf/?url=https%3A%2F%2Fwww...
The company is worth $200B at best. Where is the growth going to come from?
Claude has been pushed into every software org possible already by force. Companies will evaluate the actual benefits in 2027 and usage will go down.
There is no other product. It remains to be seen how long investment banks and influencers can spin the narrative.
And still, so called stock owners get no voting rights.
You don't get voting rights at the casino, either.
Why do they have to go with the ridiculous gazillion dollar valuations? Is it really such a bad thing to be valued at "only" 400B or whatever more realistic number, why does it have to be in the trillions?
the revised S-1 behind that $895B commitments figure says nine of the ten largest pharma cos, so it seems interesting that's what they seem to be basing off of
I have yet to see anyone releasing a nuanced report on all the metrics involved, everything I've found is short term or incomplete or misinformation, I am neutral, but the matter is complicated, they might pull it off, despite all the misleading marketing and Chinese competition and data center obstacles.
Too much sensationalist reports out there, cherry picked by all sides.
See also: Ant: Can unprecedented revenue growth outrun irresponsibly large investment? <https://news.ycombinator.com/item?id=50326281#50335632>
No wonder when they're STEALING ALL THE CONTENT.
> It reported a $104 billion net loss during the period, though most of that was due to the change in fair value of convertible notes and derivative liabilities.
Look, I've seen and I understand the arguments that the GAAP income adjustment for this kind of thing is fake, but a number of them seemed to take it for granted that it was a "one-time" adjustment. If it's going to keep happening I really feel like we ought to insist on a more comprehensive argument. The accounting principle must be generally accepted for a reason: what shenanigans could someone get up to by reporting net income without this adjustment, and how do we know Anthropic is not getting up to those shenanigans?
It’s like shareholder dilution from issuing more stock or stock options. Existing shareholders might have a smaller share of the company than they thought. The accounting is weird.
We aren’t shareholders, though, so why should we care?
>We aren’t shareholders, though, so why should we care?
They're going through an IPO, so unless this is a one-off thing, people reading this thread will become shareholders, either directly or through their pension pots.
Sometimes accounting is weird because it's necessary to prevent tricks. Why do all your held-to-maturity bonds become "tainted" and get revalued at market prices when you sell a couple, even though nothing has changed about them? Because any other rule would let a company which cannot in fact hold bonds to maturity get up to tricks and systematically overstate their assets.
So while yes, this does look a lot like shareholder dilution, that doesn't really satisfy the question of why it's not treated that way under GAAP. Perhaps GAAP is just wrong, or perhaps there's some possible tricks which Anthropic is provably not engaging in. What it's definitely not is a one-time adjustment now that it's happened in two consecutive quarters.
It's a good question: why are the GAAP rules what they are? It needs to be investigated, though.
The reason it happens in multiple quarters is likely that Anthropic's valuation has increased in multiple quarters. Every time Anthropic's valuation increases, the convertible bonds are worth more money. Other shareholders have "lost" money compared to what would have happened if Anthropic had sold regular bonds instead of convertible bonds. Except that no, the value of their shares is increasing too. An increased valuation is actually good for them.
That's how equity works. If someone owns 1% of a company and its valuation doubles, their share is worth twice as much and so is yours and we wouldn't even normally consider it to be diluting the other shareholders.
So the question is whether convertible bonds are worth more than regular bonds and the answer is, yes, absolutely, but they are partially equity, and it's the equity part that's increasing when the company's valuation is increasing.
And if the company's valuation went down then the value of the equity part of the convertible bonds would decrease. Maybe that would be a profit by GAAP rules?
> The bond market works best when there’s an end in sight, Lyman noted. Every AI investor believes there’s a light at the end of the tunnel. We just have no idea how long this tunnel is.[0]
It is like someone who loses repeatedly at Roshambo while saying: "Okay, then. Best two out of three- three out of five- four out of seven- five out of nine..."
[0] https://edition.cnn.com/2026/10/30/business/ai-stocks-bonds-...
Lol, pretty much exactly that. "Wait wait, I wasn't ready yet."
These are all supposed to convert to regular shares at or around IPO IIRC. If they're still pulling that crap after that point we'll know. There's an end date for those contracts that we're expecting
Anyone with successful portfolios here - should a retail pleb invest?
Only with money you can afford to have sit (hold) for 10+ years. The closest analogy we have dotcom bubble and crash. We don't know if we are in 1998 or 1999. For most investors who got in in 1999 and held, they didn't break even until 2007 or longer. Versus the investor that got in 1998, was break even in 2003.
Obviously the investors who bought in at the bottom did the best :-) Most people who are crying 'AI bubble' predicted 8 of the last 2 bubbles.
My personal opinion is ANT IPO is like the netscape IPO of 1995, its the start of the cycle not the end. Things are going to get really crazy over the next 3-5 years, the last year was just a warmup. I don't think people realize how much credit fiat dollars have been and can be created, and how $1T companies will be commonplace. I think humans have a mental block, where we think dollars are like rulers or other fixed measuring devices, so you can look at the same number at different periods of time and compare. credit backed currency completely warps our sense of scale when it comes to measuring things in dollars.
1T companies common place in the next 5 yrs? 20 yrs?
Do not FOMO.
It's not the same thing as early NVidia / Amazon / Google IPO: these companies didn't IPO at a value immediately placing them in the top 10 (by companies market cap) and certainly didn't IPO at values representing 5% of the US GDP.
Can it x2? x3? x4? Maybe.
Can it x100 and represent five times the US GDP? I don't think so (I know, I know, the US GDP would grow too).
I know ANT ain't SPCX, but look at what happened so far to SPCX's valuation.
Could it 1/100x? Yes.
helllll no
Release the weights instead
Half of the revenue in the first 9 months of 2026 was in Q3? No one is going to believe that. Focus on the net loss of $23.29 billion in Q3, that's more instructive.