Are we really just seeing the AI market sort itself out? We have leaders [OpenAI, Anthropic, Google]. Secondary players [xAI, Meta] once looked promising, now sort to the back.
We don't need half a dozen frontier labs. They're capital intensive.
And since neither Metas AI / metaverse investments worked out, they're back to the same business they've always had: social media. Which is an OK enough business, but not high growth.
Do a search for "Meta" in HN and their 2022 layoffs of 11000 people is one of the most-voted posts with over two thousand upvotes. Now in 2026 it barely registers as news.
Almost everyone says turmoil and bad morale is the default at Meta. The default is lots of turf-protection, project stealing, and backstabbing. It will probably get worse with these news coming out.
Recently I've also heard of teams having very little amount of time to prove themselves before they get reorg'ed - which might be one of the reasons that their progress on AI has been slow.
The only person I know that is happy at Meta is someone who works at one of the acquisitions. That specific acquisition doesn't seem to have been infected by Meta's culture too much for now.
I just hope other CEOs don't see this and get antsy about following suit.
>Meta's planned AI investments follow a series of setbacks with its Llama 4 models last year, including criticism that it provided misleading results on the benchmarks it used for early versions. It abandoned the release of the largest version of that model, called Behemoth, which had been due out in the summer.
>The superintelligence team has been working to reassert the company's standing this year by building a new model called Avocado, but the performance of that model has also lagged expectations.
20% headcount reductions in return for a 600bn capex outlay to train that next gen base model
This... can't be a signal of strength. There's a fine line between being agile and being erratic.
AI investment makes total sense as a proximal explanation. Minimize debt by trimming OpEx, then reinvest in compute. Seems smart.
And yet - this is what, the third layoff in 5 years? And weren't they doing aggressive performance cuts too? Are they workforce planning in 12 week sprints or something?
This reminds me of an overspending sports team: just toss together overpriced players/coaches, underperform, fire them all, do it again.
Are we really just seeing the AI market sort itself out? We have leaders [OpenAI, Anthropic, Google]. Secondary players [xAI, Meta] once looked promising, now sort to the back.
We don't need half a dozen frontier labs. They're capital intensive.
And since neither Metas AI / metaverse investments worked out, they're back to the same business they've always had: social media. Which is an OK enough business, but not high growth.
Do a search for "Meta" in HN and their 2022 layoffs of 11000 people is one of the most-voted posts with over two thousand upvotes. Now in 2026 it barely registers as news.
Because it hasn't happened yet.
A company in incredible turmoil. Would not want to be there now. Morale must be brutal.
Almost everyone says turmoil and bad morale is the default at Meta. The default is lots of turf-protection, project stealing, and backstabbing. It will probably get worse with these news coming out.
Recently I've also heard of teams having very little amount of time to prove themselves before they get reorg'ed - which might be one of the reasons that their progress on AI has been slow.
The only person I know that is happy at Meta is someone who works at one of the acquisitions. That specific acquisition doesn't seem to have been infected by Meta's culture too much for now.
I just hope other CEOs don't see this and get antsy about following suit.
>Meta's planned AI investments follow a series of setbacks with its Llama 4 models last year, including criticism that it provided misleading results on the benchmarks it used for early versions. It abandoned the release of the largest version of that model, called Behemoth, which had been due out in the summer.
>The superintelligence team has been working to reassert the company's standing this year by building a new model called Avocado, but the performance of that model has also lagged expectations.
20% headcount reductions in return for a 600bn capex outlay to train that next gen base model
This... can't be a signal of strength. There's a fine line between being agile and being erratic.
AI investment makes total sense as a proximal explanation. Minimize debt by trimming OpEx, then reinvest in compute. Seems smart.
And yet - this is what, the third layoff in 5 years? And weren't they doing aggressive performance cuts too? Are they workforce planning in 12 week sprints or something?
This reminds me of an overspending sports team: just toss together overpriced players/coaches, underperform, fire them all, do it again.
This will get them back to pre-covid levels.
79k * 0.8 = 63.2k
At the end of 2019 they had 45k employees.
It's more accurate to say that this will get them to late 2023 levels which was 67k employees:
https://stockanalysis.com/stocks/meta/employees/
Meta is close to achieving "AGI" internally (layoffs).