Microsoft pairs U.S. voluntary buyouts with $37.5B AI spending plan
- Microsoft began its first-ever U.S. voluntary buyout program on April 23, aimed at some older, longer-tenured workers as AI spending keeps climbing. - The offer covers about 7% of U.S. staff — roughly 8,750 people — using a “Rule of 70” age-plus-tenure cutoff. - The bigger backdrop is capital intensity: Microsoft just said 2026 capex could reach about $190 billion, mostly for AI infrastructure.
Microsoft is doing two things at once. It is offering some U.S. employees a paid exit, and it is pouring astonishing amounts of money into AI infrastructure. Those moves are not identical, but they rhyme. The company is reshaping itself for an era where the bottleneck is no longer office space or even headcount — it is compute, power, and the ability to ship AI products fast. ### What exactly did Microsoft announce? On April 23, Microsoft told employees it would run a one-time voluntary retirement program in the U.S. It is the company’s first such buyout in its 51-year history. The offer is open to workers at senior director level and below whose age plus years at Microsoft add up to at least 70. People on sales incentive plans are excluded, and eligible workers were told to expect details on May 7. About 7% of Microsoft’s U.S. workforce qualifies — roughly 8,750 people out of 125,000 U.S. employees. (cnbc.com) ### Why that “Rule of 70” matters This is not a broad early-retirement wave for everyone. It is a narrow filter aimed at older, longer-tenured employees who are still below the very top of the org chart. Basically, Microsoft gets a way to reduce payroll and open roles without the uglier optics of a(cnbc.com)makes clear this is targeted workforce shaping, not a general perk. (cnbc.com) ### Is this really about AI? Not in the simple “AI replaces these exact jobs” sense. But yes, AI is the backdrop. Microsoft has been ramping data-center spending to serve cloud customers and its own generative AI products, and management just told investors to expect roughly $190 billion in capital(cnbc.com)mpany is spending that much on GPUs, servers, networking gear, and facilities, every other cost line gets scrutinized harder. (microsoft.com) ### Where did the $37.5 billion number come from? That figure appears to come from earnings-preview estimates, not from Microsoft saying “we committed $37.5 billion” in one clean announcement. Microsoft’s actual latest disclosure was even bigger on the forward view — the company said capex for calendar 2026 could be about(microsoft.com)re Microsoft reaffirmed an enormous AI infrastructure buildout, not after a single $37.5 billion headline pledge. (computing.net) ### Why do buyouts instead of layoffs? Buyouts are gentler, and they can be cheaper politically even if they cost cash upfront. They also let a company target expensive, long-tenured workers without publicly saying that is what it is doing. Microsoft already cut jobs in multiple roun(computing.net)d AI-heavy priorities. (cnbc.com) ### Does this mean Microsoft is in trouble? No — if anything, the weird part is that Microsoft is doing this from a position of strength. The company is still growing, Azure demand is strong, and investors are mostly focused on whether Microsoft can turn giant AI spending into durable revenue. The (cnbc.com)egacy roles matters less than building enough compute capacity to stay ahead. (microsoft.com) ### What should readers take from it? The cleanest read is that big tech’s AI transition is now showing up in org charts, not just product demos and earnings slides. Microsoft is not simply adding AI on top of the old company. It is funding a new capital-heavy model and quietly making room for it. (cnbc.com)
