Xbox says it's growing again. The people it fired might disagree.

On October 6, 2026, Xbox CEO Asha Sharma told an all-hands town hall that the worst was behind the division. "We're making the right decisions to reinvest in the business," she said. "After an all-time low, we've started to return to growth, more than two to three times what we were, and we expect that continue." ^[1][4]

We know this only from a transcript obtained by Windows Central, which says trusted sources verified it. Microsoft has not published the transcript ^[4]. Windows Central later clarified that Sharma was pointing at a chart of first-party growth specifically, meaning games Xbox owns and publishes, not the whole business ^[4]. This story is still developing, and everything about the town hall rests on that single leak.

Here is the paper's reading. The claim may well be true. But it answers a different question from the one the layoffs answer. Growth is a statement about revenue and play. The cuts are a statement about cost. Sharma's own July memo framed Xbox's problem as margins, and said the business "did not grow with our audience" even as 200 million new players arrived ^[5][8]. So a recovery chart cannot serve as a verdict on the restructuring that preceded it, and a company that lets the two blur together is asking its staff and its audience to forgive the cuts on the strength of a number it has not shown them.

What she said, and what we cannot check

Sharma told staff that at the start of the year Xbox was declining in hours played, player numbers and revenue. Hours played, she said, have now stabilised, and first-party has returned to growth after year-over-year declines ^[1][4]. She named three recent releases with "strong" starts: Gears of War: E-Day, Minecraft Dungeons 2 and World of Warcraft: Forever. She also said positive brand sentiment had risen 30% over the past year, and that the old Game Pass strategy had sent the service's profitability into "freefall" before it was repriced and stripped of day-one Call of Duty ^[1][4].

Read closely, the claim is narrower than the quote sounds. Hours played were described as stabilised, not growing. Player numbers and revenue were not said to have turned ^[4]. "Two to three times what we were" has no stated base period, no absolute figure and no named metric, and a multiple means little when it is measured from what Sharma herself called an all-time low. Windows Central reported earlier that Xbox declined for four straight quarters in fiscal 2026, with Microsoft insiders blaming Call of Duty: Black Ops 7 and a thin slate of major releases ^[8]. After four quarters like that, a rebound is almost the default. The question is how big it is, and nobody outside the meeting can say.

The timing also deserves notice. In July, Sharma promised, "We will return to growth in 2027" ^[5]. On July 29 she wrote that Xbox expected "to return to growth by the end of FY27," and Satya Nadella told investors the same ^[8]. Three months later she tells staff it has started. That may simply be good news arriving early. It also means a promise made to investors is being described as under way in a room, in words that no auditor has seen.

What it took

The July 6 memo, "Resetting Xbox," called the plan "the most significant restructure in XBOX history." It cut about 3,200 Xbox roles across fiscal 2027, about 1,600 of them that day, and sent four studios out to new management ^[5]. The cuts sat inside a wider Microsoft round of about 4,800 jobs, roughly 2% of the company's global headcount ^[6]. Management layers fell from as many as 14 to no more than five, and vendor spend was halved ^[5][6].

The reasoning was blunt, and it was Sharma's own. "Our business today is not healthy," she wrote. "We are operating at margins that are 3-10x lower than comparable platform and publishing businesses." In a typical year, she added, Xbox "lost 64 cents for every dollar we invested" in its studios ^[5]. GeekWire noted that Nadella has argued the era of subsidising Xbox as a living-room strategic bet is over ^[6].

On September 22, Matt Booty announced another 268 role eliminations across Halo Studios and other first-party studios, and said the company was "roughly three-quarters of the way through previously announced restructuring" ^[3]. Staff posts collected by Game Developer indicated cuts at Rare, Treyarch, World's Edge, Activision and The Coalition ^[7]. Eurogamer counts more than 1,800 job losses so far and says the process is not expected to finish until next summer ^[1]. The announced plan is 3,200 by the end of the fiscal year ^[5], so well over a thousand more are still to come.

The studios are where this lands hardest. Compulsion and Double Fine went back to management in August with their IP, catalogues and runway funding ^[3]. Two separate agreements for Ninja Theory, the Hellblade studio, fell through, so consultation on a proposed closure will begin ^[3]. Consultation at Arkane continues through the end of the year ^[3]. Whatever the town hall chart shows, it does not show Ninja Theory.

The view from Seattle

The clearest account of what the cuts mean for the people on the other side came the same week from Philip Holt, who runs Undead Labs. The State of Decay studio was bought out of Microsoft in September by its own staff, through a new holding company called Full Measure, after a deal with unnamed buyers fell through ^[2]. Rock Paper Shotgun reported, citing The Game Business, that the buyer was Tencent-owned UK studio Splash Damage ^[2]. That attribution is the outlets' and not confirmed here.

Then came the part that makes the story hard to file under rescue. Layoffs followed within a day of the buyout being confirmed ^[2][7]. The studio said it was becoming an "independent, employee-owned studio" and laying off a "significant number" of employees, some of whom worked on State of Decay 3. It gave no figure. "These are people we'd hire again in a heartbeat," it said, "and we hope the industry takes notice." ^[7] Staff bought the studio, and some of the staff did not get to stay in it.

Holt explained the pressure in an interview with The Game Business, relayed by Rock Paper Shotgun ^[2]. "There's capital available," he said, "but I think it's nervous, perhaps sitting on the sidelines waiting for obvious slam dunks." The sticking point, he said, was "the amount of money and cost of team," and the speed of decisions. Investors scrutinised the studio's unit labour cost because it sits in Seattle, which he called "one of the most expensive places in the world to make video games," while games arrive from places where labour costs far less. He described a "race to the bottom on the economics of labor" ^[2]. Saber Interactive CEO Tim Willits has publicly urged publishers to buy teams in lower-wage regions, naming South America ^[2].

Set Holt next to Sharma and the two are working the same sum from opposite ends. She cut 3,200 roles because a dollar invested in studios was coming back as 36 cents ^[5]. He found cautious money that scrutinised what a Seattle team costs ^[2]. Both are saying that the work of making games is priced higher than the market will pay for it. The recovery narrative, and the reasons for it, belong to that sum.

The strongest objection

The best case against this reading goes as follows. Growth is growth. Xbox did decline for four quarters, Game Pass was bleeding, and a CEO who inherited that in February cannot be blamed for telling her staff when the curve bends. The cuts were never designed to raise hours played. They were designed to fix margins, and nothing in the town hall remarks says otherwise. Demanding that she hold back good news until the human cost is tallied would be asking for a different job.

Most of that holds. Three named releases landed well. Hours played stopped falling. I do not think the improvement is fake, and nothing in the dossier suggests it is. It would also be unfair to say the cuts produced the growth. A shorter org chart does not sell a Gears game.

But the objection answers a charge I am not making. My complaint is about what the sentence is allowed to do. "We've started to return to growth" arrives at an internal meeting, with the reset three-quarters done and a quarter still to go, among people who have watched colleagues leave ^[3]. In that setting it works as a reassurance that the pain was justified, and only a number could support that. The number has not been published, and the metric is first-party only ^[4]. The nearest thing to a verifiable figure is the 30% rise in brand sentiment, and that is about feeling, not money. Executives are entitled to good news. They are not entitled to have it graded on a curve nobody else can see.

There is also a plainer point about how companies keep their own history. A reset memo is a record, and so is a town hall transcript, but only one of them was published by the company. The July memo is on Xbox Wire, signed and dated ^[5]. The October claim exists because someone leaked it ^[4]. A turnaround story that lives only in a leak will be remembered as whatever the next memo says it was.

What this means

Xbox is probably better off than it was in the spring, and it is certainly smaller. Both things are true, and the second was the plan. The risk is that the first gets used to launder the second, so that "return to growth" becomes the whole account of a year in which more than 1,800 people, and eventually perhaps 3,200, lost jobs, one studio faces closure and another bought itself out of Microsoft and then cut its own colleagues.

Holt's warning is the one to keep. If investors want obvious slam dunks and the measure of a team is its unit cost, then the savings Sharma banks are not a private matter for Xbox. They are a price signal to every studio still hunting for money. The people Microsoft let go are part of that market now, and some of them will be rehired by someone who prices them differently.

What to watch

First, whether Xbox publishes anything that lets outsiders check the growth claim, since the leaked transcript is the only record so far. Second, what the final quarter of the reset means for Ninja Theory and Arkane. Third, whether the promise of growth by the end of fiscal 2027 is met on reported financials rather than an internal chart. Fourth, whether Full Measure and the other divested studios keep their people once runway funding runs out, and whether Holt's sidelined investors ever step in. Undead Labs says State of Decay 3 is still aimed at next year, with a day-one Game Pass release under a new publisher ^[2][7]. That game, and who is left to finish it, will say more about this recovery than the chart did.

In July, Sharma wrote that "history is full of companies that mistake longevity for inevitability" ^[5]. Surviving a reset is not the same as having won one, and the people who know that best are the ones who are no longer on the payroll.