Meta reopens manager roles in the AI division it staffed by reassigning 7,000 people

Meta's case for flattening was that fewer layers would improve decision-making, cut bureaucracy and help offset the rising cost of its AI investments ^[15]. The second quarter tests that case. Revenue of $60.8 billion, up 28% from a year earlier ^[9], implies about $47.5 billion in the year-ago quarter and $13.3 billion of new revenue ^[1]. Total expenses of $42 billion, up 55% ^[10], imply about $27.1 billion a year earlier and $14.9 billion of new cost ^[2]. Expense dollars grew about $1.6 billion faster than revenue dollars, so operating income fell from roughly $20.4 billion to $18.8 billion, and the operating margin from about 43% to 31% ^[3]. Fortune reported the expense line also absorbed costs tied to the May reductions ^[10].

The headcount has not landed yet. Meta closed the quarter with 75,472 employees, down 3% sequentially, and that count still includes the roughly 8,000 people affected by the May cuts ^[8]. A 3% decline implies about 77,800 the quarter before, a sequential fall of only about 2,300 ^[4]. Take the 8,000 out and the trailing figure is nearer 67,500, about 13% below the prior quarter ^[5]. The May reduction was described as about 10% of the workforce, which puts the pre-cut base near 80,000 ^[7][6]. Add the 6,000 open positions Meta scrapped and the initiative removed 14,000 positions against plan, about 18% of that base ^[7][8].

Against those numbers, the reversal is small. Roughly 7,000 employees were moved into Applied AI earlier this year, about one in eleven of quarter-end headcount ^[4][7]. Those individual contributors are the ones now being asked whether they want to manage again, voluntarily ^[1][2]. Meta has not said how many manager slots are on offer, and it did not respond to Fortune's request for comment ^[5].

The reason is inferred. Fortune wrote that the decision "highlights the challenge companies face in balancing efficiency, rapid AI development, and workforce coordination" ^[12]. What Zuckerberg said on the second-quarter call was about speed: "I'm also excited about how AI is helping our teams speed up product development," he said ^[11].

This is the second pass through the same manoeuvre. In 2023, during the "year of efficiency", Meta asked many managers and directors to take individual contributor jobs or leave, in a process it internally called "flattening" ^[6]. The newer reorganisation has drawn complaints: Wired reported employee frustration over the Applied AI rollout, and some reassigned workers were later given the option to look for other roles inside the company ^[13].

The record fits more than one reading. It could be a narrow correction in one division, cheap enough that nothing follows from it. It could be an early signal that the May cuts went past the point where 7,000 reassigned engineers coordinate themselves. If so, the layer returns across product groups and shows up in expenses a quarter or two before it shows up in revenue. Or it is reorganisation churn, where the cost lands as disruption and never appears in payroll. I'd take the first, with a lean toward the second in the teams closest to shipping models. If the offer travels outside Applied AI, or if third-quarter headcount comes in flat once the 8,000 are off the books, the narrow reading is wrong.