Microsoft Offers Buyouts to About 60,000 Employees Amid AI Data Center Push
Microsoft offered buyouts to about 60,000 employees, a quarter of its workforce, to free funds for AI data centers.
The employees reportedly have a score of 70 or below on Microsoft’s internal experience scale, meaning the most senior employees affected are vice presidents. Workers have until early December to accept the offers and will leave later that month.
Some marketing, sales and engineering teams have reportedly decided not to participate. It is unclear which specific business units will see cuts. SiliconANGLE reported that Microsoft’s loss-making Xbox video game business seems likely to be among the affected groups; after Microsoft let go 1,600 employees in the division in June, GeekWire reported that additional layoffs were expected.
According to Business Insider, workers with a score of 64 or below will receive one week of pay for every six months spent at the company, while staff at levels 65 to 67 will receive two weeks. The package also includes stock vesting for six to 12 months, continued healthcare coverage for a limited time and outplacement support, according to the report.
Microsoft Chief People Officer Amy Coleman wrote in a memo to employees: “The Voluntary Exit Offer is new for us. We’ll learn from it. At the same time, we’ll keep helping people focus on the work that matters most, build new skills and, where it makes sense, find new opportunities across Microsoft.”
The buyouts come after Microsoft reported revenue growth of 18% year over year in the three months ended June, to $90.01 billion, and adjusted earnings of $4.74 per share that beat the consensus estimate by more than 12%.
The company opened 31 new data centers last quarter, more than half the total number of cloud facilities it had at the start of the year. SiliconANGLE reported that Microsoft expects 2026 capital expenditures to be down from its April projection of $190 billion but still significantly above 2025 spending.
Earlier this month, Microsoft introduced an inference accelerator called Maia 300, which it says delivers 40% more tokens per megawatt than leading graphics cards. The company said it plans to make the Maia series its primary source of inference capacity.