Meta planning layoffs as AI costs mount(reuters.com)
reuters.com
Meta planning layoffs as AI costs mount
https://www.reuters.com/business/world-at-work/meta-planning-sweeping-layoffs-ai-costs-mount-2026-03-14/
22 comments
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.
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.
It was always rough, but this is by far the worst I've ever seen or heard. The thrash and org maneuvering are feral.
>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
>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
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.
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.
> they're back to the same business they've always had: social media. Which is an OK enough business, but not high growth.
Amazing that Meta is even mentioned in the same breath as Google considering the ocean sized difference in what they do and the services they offer.
I know the reason is market cap. But that's the amazing part. That a company as limp and unimpressive as Meta could have over a trillion dollars in market cap.
Amazing that Meta is even mentioned in the same breath as Google considering the ocean sized difference in what they do and the services they offer.
I know the reason is market cap. But that's the amazing part. That a company as limp and unimpressive as Meta could have over a trillion dollars in market cap.
They are quite similar in the sense that both generate most of their revenue through ads.
Right but Google has loads of market leading services they use to deliver those ads, and Facebook has a crappy dated website and dumb ideas that lose them money which they brand the company on.
I know I'm simplifying, but not by much imo.
I know I'm simplifying, but not by much imo.
They have pretty similar P/E (currently Meta a bit better).
Watching ads on facebook/instagram is addictive/profitable.
Watching ads on facebook/instagram is addictive/profitable.
> They have pretty similar P/E
That's the baffling part. I would think YouTube alone would crush Facebook in profitability. Then add all the other things Google does and it's not even close.
I understand the reality, but I don't get it.
That's the baffling part. I would think YouTube alone would crush Facebook in profitability. Then add all the other things Google does and it's not even close.
I understand the reality, but I don't get it.
I'm reminded of the Hastings quote about Netflix's biggest competitor being sleep. As far as the various infinite-scrolling ad-machines are concerned, the primary input is users' time, and whatever the service does or looks like is just window dressing in service to triggering dopamine hits to keep users scrolling.
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.
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 is what, the third layoff in 5 years?
They have done more targeted layoffs as well like the FAIR layoffs in 2025 and the RL layoffs in January 2026, as well as the 5% performance cut in February 2025.
They have done more targeted layoffs as well like the FAIR layoffs in 2025 and the RL layoffs in January 2026, as well as the 5% performance cut in February 2025.
Meta is close to achieving "AGI" internally (layoffs).
> achieving "AGI" internally
There comes an uncanny point where a Corporation's beleaguered flailing is indistinguishable from Actualised General Incompetence.
There comes an uncanny point where a Corporation's beleaguered flailing is indistinguishable from Actualised General Incompetence.
This is fantastic. It does beg the question for me of how much incompetence is acceptable from a purported (traditional) AGI system before it no longer qualifies as such. After all, the competence of humans varies wildly.
On the last earnings call Zuck empathized how much revenue is being allocated to salaries and RSUs. I take those types of comments during earnings calls to be signals to investors that cuts will be coming to bring expenses down by the next quarter. How many rounds between earnings is speculation, but you know at at least one big one.
A friend who works in a growing division within Meta also said last week they are re-orging again.. this time flattening out everything and making first level SDMs go back to IC. It’s always been a cut throat final-stage capitalism place to work. Nothing they do should be surprising.
A friend who works in a growing division within Meta also said last week they are re-orging again.. this time flattening out everything and making first level SDMs go back to IC. It’s always been a cut throat final-stage capitalism place to work. Nothing they do should be surprising.
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/
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/
Late 2023 levels was after the November 2022 layoffs as well (~11k cut).