Microsoft shed 4,800 roles on July 6, a touch over 2% of its workforce, but the headline figure barely explains why the stock has surrendered roughly $1.2 trillion in market value since its October peak. For anyone with an S&P 500 index fund in a 401(k), about four dollars of every hundred is tied to this one name, so the retreat matters whether you track the company or not. The market’s anxiety is not really about headcount. It is about the $190 billion in capital spending Microsoft has earmarked for this fiscal year, the vast bulk of it flowing into AI data centers, and whether those outlays will ever earn their keep.
The cuts themselves tell a more pedestrian story. Roughly 1,600 came from Xbox, where chief executive Asha Sharma told staff the division runs at margins “3-10x lower than comparable platform and publishing businesses.” Microsoft is also spinning off four game studios to stand alone. Most of the remainder hit commercial sales and consulting. Chief People Officer Amy Coleman insisted the eliminated roles are not being directly replaced by AI, even as she acknowledged the technology is changing how work gets done. Brad Smith, the president and vice chair, put it more bluntly: Microsoft can only be a strong employer if it has a successful business.
That success is now a debate about math. Chief Financial Officer Amy Hood told analysts in April that expenses would climb about 7% in the June quarter, yet she still guided for margin expansion in fiscal 2026, after absorbing roughly $900 million in one-time retirement costs. The company ran its first voluntary retirement program this spring and summer; about 30% of the 8,750 eligible U.S. employees took it, softening the blow of involuntary separations. The visa filings that stirred outrage, 2,879 H-1B labor condition applications this fiscal year, are a sideshow. The real variable is whether AI capacity turns into revenue faster than it turns into depreciation.
Xbox is the canary. If a platform business with a locked-in user base and recurring subscription revenue cannot crack margins within shouting distance of peers, the return on hundreds of billions of silicon and electricity becomes a fair question. The studio spin-offs suggest Microsoft is willing to shrink the surface area of its bets. But the core wager, that AI infrastructure demand justifies a capex run rate unseen in the company’s history, remains unproven.
What to watch next is not the next round of layoffs. It is the fiscal 2026 margin trajectory Hood promised, the pace at which AI workloads fill the new data centers, and whether Xbox can prove its model is fixable or merely subscale. The market has already priced in the ambition. Now it is waiting for the economics to catch up.