The Xbox collapse was a platform identity crisis
Stratechery founder Ben Thompson argues Microsoft tried to run a software and service business on top of a console that kept losing money, a contradiction he says is shaping AI competition.
Stratechery founder Ben Thompson argues Microsoft tried to run a software and service business on top of a console that kept losing money, a contradiction he says is shaping AI competition.
Stratechery founder Ben Thompson used this month's TBPN episode to name what Microsoft spent a decade avoiding. On "Ben Thompson Explains the XBOX Disaster", Thompson argues Xbox collapsed because Microsoft tried to run a software-and-service business on top of a console that kept losing money, and the company never closed the gap between the two models.
That diagnosis arrived the same week Microsoft cut thousands of jobs and shut five studios in what TechTimes called the largest gaming layoff in years. GameRant's read on the restructure reaches the same conclusion from a different angle: Game Pass economics is the load-bearing failure. Subscription revenue was supposed to convert a cyclical, hardware-dependent business into a recurring platform, and AAA development costs rose faster than the service could monetize them. Microsoft kept funding the loss with the balance sheet.
Thompson's Game Pass diagnosis has been the subtext of Xbox coverage for years. This month it became the text, when Xbox CEO Asha Sharma admitted Microsoft's gaming strategy "failed to work out". Sharma was speaking about Game Pass specifically, but Thompson reads the admission as Microsoft finally naming the failure it had been working around: the platform identity the company bet on never worked.
The $69 billion Activision Blizzard acquisition was the escape hatch Microsoft hoped would solve the Game Pass math. Thompson treats it as a hedge against a structural problem no amount of content could fix. Acquiring Call of Duty and the rest of Activision's catalog lowered Microsoft's per-subscriber content cost, but it locked the company into hardware losses for another decade. The deal also produced the kind of regulator fight Thompson has flagged before. Polygon's running tracker of Xbox's upheaval is the documentary record of what that timing produced.
Thompson's GTA 6 thesis, that the game should cost $200 at launch, makes the same point from a different direction. In his framing, AAA development costs have outrun the $70 price the market currently accepts, and publishers like Take-Two will need to charge more if they want to fund the next generation of games without leaning on subscription services that cannot pay for them.
Thompson argues regulators often block the deals that would have prevented exactly these restructures. The Federal Trade Commission's campaign against the Activision merger is the clearest recent example. By trying to stop the deal that would have insulated Microsoft from rising development costs, regulators ensured Microsoft would face those rising costs without the content library the merger would have delivered. Xbox's layoffs are the visible cost of that timing.
The pattern Thompson uses Xbox to illustrate is not specific to gaming. In his view, AI is the live version of the same fight. OpenAI, Anthropic, and Google are spending on training and inference at rates current API and subscription pricing cannot sustain. The companies that figure out how to charge for AI the way Microsoft tried to charge for Game Pass will face the same identity question: are you a hardware business, a software business, or a service, and which revenue model funds which cost. Microsoft's Xbox outcome is a preview of what happens when a company picks the wrong answer.
What Microsoft does next is the open question. The post-restructuring Xbox is meant to be smaller and software-led, but the studios that fed Game Pass are mostly the studios that just got cut. Thompson does not promise a recovery. He argues the diagnosis was visible years ago, and the cost of missing it is now on the company's income statement.