Microsoft Offered 60,000 Employees the Door. It Spent $45 Billion on Servers.

The offer went to more than 60,000 people, roughly a quarter of the staff, and it came with terms unusual for a company cutting costs: severance, stock that keeps vesting, health coverage that continues. Nobody had to take it. That is the whole design. A voluntary exit is a resignation the company pays for, and it lets the same reduction in headcount be described as a choice.

Anyone who works at a large company should read the two numbers together. In the same stretch in which Microsoft asked a quarter of its people to consider leaving, it spent $45.2 billion on capital projects in a single quarter and told investors to expect more than $50 billion in the next one, most of it on chips, servers and data centers. Then it signed a 20-year agreement to buy nuclear power.

The power contract is the detail that explains the rest. Twenty years is longer than most of Microsoft’s product cycles and longer than most of its customers’ planning horizons. A company does not commit to a generating source for two decades because it expects to need less electricity in 2031. It commits because the thing it is building consumes power at a scale that only a long contract can secure.

What is being bought with that money is compute, and the logic behind the trade is a reversal of the one that ran human resources departments for twenty years. The scarce input used to be people, and the strategy was to hire the best ones before a rival did. The scarce input now is capacity: chips that are allocated rather than sold, sites with power and water, and machines that can be kept busy. Microsoft’s own framing is a shift from competing for talent to competing for compute.

That shift lands hardest on the middle. Coding assistants that write a working draft of a routine function remove the junior work that used to be how a company trained its next layer of managers. If a team of eight can now do what a team of twelve did, the eight do not need as many people coordinating them, and the coordinator role is exactly the one that disappears quietly rather than in an announcement.

The retention test that follows is blunt. You stay if you bring in revenue or if you keep critical infrastructure running. Everything between those two poles, the reporting, the internal tooling, the meetings that exist to align other meetings, is a candidate for automation, consolidation or the exit offer.

There is a real case for the program as a humane instrument. A layoff is decided in a room and delivered in a fifteen-minute conversation, and the person on the other side has no say in the timing. An exit offer with severance, vesting stock and continued health coverage hands the decision to the employee, who may have been thinking about leaving anyway and now gets paid to do it. Companies that ran similar programs in past downturns found takers among exactly the people they would have struggled to lay off, including long-tenured staff near retirement.

The case against is older than the technology. Voluntary programs tend to remove the people who can most easily find another job, which is not always the same set as the people a company can spare. Institutional knowledge sits in the middle layer that this kind of offer targets, in the person who knows why a system is configured the way it is and which customer will call if it changes. That knowledge does not appear on a balance sheet and does not come back when the cycle turns.

The capex number carries its own risk. Fifty billion dollars a quarter is a bet that demand for AI capacity keeps growing faster than the depreciation schedule on the buildings that house it. If that bet is wrong, the correction does not arrive as a smaller bonus. It arrives as idle capacity and a second round of cuts, this one without the word voluntary attached.

For the 60,000 who received the offer, the arithmetic is personal and unglamorous: the severance in months, the vesting schedule, the cost of replacing the health plan, the odds of finding comparable work in a market where every large employer is running a version of the same program. Some will take it gladly. Some will stay and discover that the team they return to is half its old size.

Microsoft has not described the program as a retreat from hiring, and on the evidence of the capital budget it is not one. It is a reallocation, from payroll to plant, announced in the language of choice because that is the language that costs least. The envelope on the desk says the decision is yours. The $45 billion says where the company has already decided to put its money.

Related Posts

  • November 20, 2026
  • 16 views
Sandisk Completes Eight-for-One Split as NAND Outlook Darkens

Eight for one. Sandisk’s stock split took effect on Nov. 19, and the shares began trading on the adjusted basis on Nov. 20, a change that divides the price by…

  • November 20, 2026
  • 13 views
Qualcomm Cuts About 500 Jobs in California While It Pushes Into AI

The confirmation came on Nov 20, in the flat language companies use for this kind of news: about 500 positions eliminated in California, spread across departments in San Diego and…