Here in Singapore, some of us were up at 4 am to read the emails that made it official. Meta has been moving fast on this reorganisation all week. We’re talking about 10% of the workforce going by the wayside, while others are being put to tasks where the company thinks it can find some real growth.
If you look at the internal word, these layoff notices are being doled out in three waves, right around 4 am local time. In North America, they had to be told to stay home for the day of the announcement. They’re also putting a stop to 6,000 open positions as part of this reset.
Here are the key developments to note this week:
– Layoffs staged at around 4 am local time
– About 8,000 roles cut globally
– 7,000 staff moved to AI teams
– 6,000 open roles shut
– US severance: 16 weeks plus tenure
Competitive stakes and investor reading
You have to see it as a way to get ahead of the likes of Google and OpenAI. The numbers back it up: we’re looking at $125-145 billion in capital spending by 2026 to build out the kind of AI infrastructure it takes to keep up.
Janelle Gale, our head of HR, let staff know we’re going with a flatter model from here on out. Fewer managers, smaller pods that can move quicker. The thinking is that you can’t just put AI on top of your products; you have to build them from the ground up with it in mind.
It’s a worldwide thing. After Singapore, it was Europe and then the US. Engineering and product are where you’ll feel it most, and there may be more to come down the line.
Who is affected and how severance works
In the US, if you’re one of those leaving, you’re looking at 16 weeks of base pay, with two more for every year you’ve put in. You and your family also get 18 months of health coverage. Over in other parts of the world, the terms are a bit different.
Inside the HR memo
Gale put it in a memo as the next step in a long process. The gist of it: we want to ship faster with less red tape. She even put in a note about how hard it is to put a number on the ‘ideal’ size of the workforce when AI is changing the rules.
Some of the talk among employees has been about privacy, like with the mouse-tracking for AI training. Meta won’t go into details on that, but you can tell it’s a source of some friction. When you automate this much, the culture takes a hit.
This is no half-measure. To put 7,000 in AI and trim 8,000, not to mention 6,000 unfilled spots, is to say you don’t need as many generalists. You want specialists.
The money tells the story. With that 2026 capex plan, Meta is making its intentions known. But then again, when leadership says they can’t quite define the right headcount, you know it’s a fluid situation.
What employees should expect next
Now that the notifications are done, the teams are being put back together in these smaller units. The aim is to get AI features out the door in a hurry. Meta is willing to give up some of its breadth for the sake of focus. Whether they can pull it off, with all the questions and changes in the air, remains to be seen.
