Microsoft Cut an Xbox That Was Still Growing

On 6 July 2026 Microsoft did two things at once. It cut about 4,800 jobs, roughly 2.1% of its workforce, and it began pulling apart the game-studio empire it had spent years assembling. Ninja Theory and Undead Labs were handed to new owners. Compulsion Games and Double Fine Productions were returned to independence. Arkane, the studio behind Dishonored, entered a formal consultation with its works councilworks councilAn elected employee body that, under French and other European labor laws, a company must formally consult before major decisions like restructuring or a sale. in France over its future. Inside Xbox, about 3,200 roles, roughly 20% of the division, will be cut through fiscal 2027, with the first ~1,600 gone immediately.

A crucial and widely muddled point: these are not Activision studios. Ninja Theory, Undead Labs and Compulsion were bought in 2018, Double Fine in 2019, and Arkane arrived through the separate $7.5B ZeniMax deal that closed in 2021. All of that predates the $68.7B Activision Blizzard acquisition, which closed in October 2023. Microsoft is keeping the Activision catalog, Call of Duty, Warcraft, Candy Crush, intact. It is cutting the studios it built first, not the ones it bought last.

The framing came from Xbox CEO Asha Sharma, who did not soften it. The gaming business, she wrote to staff, is 'not healthy.' It operates 'at margins that are 3-10x lower than comparable platform and publishing businesses,' and in a typical year, she said, it 'lost 64 cents for every dollar' invested. Margin, the share of each sales dollar a business keeps as profit, is the word to hold onto. It is the hinge of the whole story.

Here is the narrow question, because a broad one would be useless. Is Microsoft cutting Xbox because gaming got worse, or because AI raised the price of holding it?

Those sound similar but they are opposite diagnoses. In the first, the patient is sick: the pieces being shed were losing money, and Microsoft is cutting rot. In the second, the patient is fine but the neighborhood got richer: the '3-10x lower margin' is not a verdict that Xbox fell, but that the bar it is measured against, Microsoft's own AI and cloud business, rose. Same layoffs, entirely different story about what AI is doing to corporate America.

Is Microsoft cutting Xbox because gaming got worse, or because AI raised the price of holding it?

Start with the claim you can actually test. If gaming were collapsing, the money-making part of it, the software and subscriptions, not the plastic boxes, would be shrinking. So we split the segment in two.

Xbox revenue comes in two very different flavors. 'Content and servicesContent and servicesMicrosoft's gaming software and subscription revenue, game sales, in-game spending, and Game Pass, as opposed to selling console hardware.' is games, in-game spending, and Game PassGame PassXbox's Netflix-style subscription: a monthly fee for access to a rotating library of games instead of buying each one., the Netflix-style subscription. That is where the margin lives, because software costs almost nothing to copy. 'Hardware' is the console itself, which Microsoft has historically sold near cost to get the box into your living room. When you pull Microsoft's own numbers apart, the two legs point in opposite directions.

In the most recent clean full-quarter split, the quarter that ended June 2025, total gaming revenue grew about 10% year-over-year, content and services grew about 13% (driven by first-party games and Game Pass), and hardware fell about 22%. Xbox hardware has now declined for four straight fiscal years. In plain terms: the profitable engine was accelerating and only the low-margin box was in decline. That is not the fingerprint of a business collapsing. It is the fingerprint of a growing software business bolted to a shrinking hardware business, and Microsoft chose to shrink the whole thing anyway.

The word 'margin' in Sharma's quote is doing quiet, heavy lifting. She did not say gaming brings in no money. She said it earns 3-10x less than 'comparable' businesses, and, separately, that the return on money invested in it has been badly negative. The first is a comparison, and a comparison depends entirely on what you compare against.

Set the numbers side by side, with a warning: they are not on one clean ruler. Secondary reporting pegs Xbox's operating marginoperating marginThe share of every dollar of sales a business keeps as profit after its running costs; a '30% margin' means 30 cents of profit per dollar of revenue. near 3% and a 'comparable platform' business near 30%. Those specific figures are outside estimates, not a Microsoft disclosure, so treat them as illustration. But one number is Microsoft's own: its Cloud gross margingross marginRevenue left after the direct cost of delivering a product or service, before overhead; a high gross margin means the core activity is very profitable. It is a different, higher measure than operating margin, so the two are not strictly comparable. sat near 69% in its FY2025 filing. Gross margin is a different, higher measure than operating margin, so this is not a strict apples-to-apples ranking. What it does show is unmistakable, the richer neighbor, and where the freed-up capital is going.

This is the intellectual center of the story. A 3% business is only 'unhealthy' relative to something better on offer. For most of Xbox's 25-year life there was no in-house alternative earning multiples of 30%, so a thin-margin gaming arm was tolerated as a strategic foothold in living rooms. What changed in 2026 is not that Xbox suddenly fell through the floor. It is that Microsoft now has an AI-and-cloud alternative that earns far more per dollar, and against that yardstick, a gap the company lived with for two decades became intolerable.

Every company keeps a mental bar, the hurdle ratehurdle rateThe minimum return a company demands before it will put money into a project; if a business earns less than the hurdle rate, its capital is better spent elsewhere., that a use of money must clear to be worth it. Raise the return available elsewhere and you raise that bar, and businesses that were fine yesterday suddenly look like a poor use of capital today. Nothing about them has to have changed. The alternative got better.

For Microsoft in 2026, the alternative is AI infrastructure, and its scale is the tell. Hold this next comparison loosely, the figures are measured on somewhat different bases, but the order of magnitude is the point.

Microsoft's stated plan for AI-enabled data centers in fiscal 2025 was about $80B. Its total capital spending is tracking toward roughly $190B for calendar 2026, and it spent $30.9B in the single quarter ending March 2026. Any one of those annual figures is larger than the $68.7B Microsoft paid for Activision, the biggest acquisition in its history. When one year of a new priority costs more than your largest-ever purchase, the math on everything else gets re-checked. That is the hurdle rate at work.

A good thesis names the number that would kill it. This one has a clean falsifier: if Xbox's content-and-services revenue turns negative year-over-year in the coming quarters, the 'still-growing software engine' claim breaks, and the story tips back toward genuine decline rather than reallocation. The same would be true if the divested studios turn out to have been the specific loss-makers dragging the segment down, in which case Microsoft was cutting rot after all, not reallocating from health.

It is worth resisting the tidy Nokia analogy here. Microsoft famously wrote down its Nokia phone business in the mid-2010s after a cloud-first strategy redefined what its capital was for. But Nokia's phone unit was loss-making and shrinking; that fits a 'cut the rot' story. Xbox's software leg was growing, and Microsoft is keeping the Activision catalog. The precedent rhymes on candor and scale, but it breaks on the very fact that makes this case interesting: the part being cut was not dying.

The honest answer is both/and. Xbox has a real problem, a shrinking, low-margin hardware business and, by Sharma's own account, years of poor returns on the money poured in. And it is being reallocated against AI, its tolerable-for-decades margin gap turned intolerable by a richer internal alternative. Both are true at once, and pretending otherwise flattens the story.

What is genuinely new is not a struggling console maker trimming costs, that is an old story. It is a mega-cap visibly shrinking a growing consumer software franchise, shedding the studios it spent years building, to keep feeding an AI build-out that now costs more per year than its largest acquisition ever did, and saying the quiet part, about margins and comparisons, out loud. The hurdle rate did not just come for Xbox. It is coming for every business inside a company with something better to spend on.