Sora Made $2.14M. OpenAI Is Worth $730B.

OpenAI shut down Sora on Tuesday. Disney walked away from a deal that had been announced as a licensing partnership and, according to Variety, included plans for the media conglomerate to take a $1 billion stake in OpenAI. The stake structure was never fully specified — whether as new equity, secondary shares, or some combination with the licensing agreement — and neither party commented on the financial terms beyond a brief statement. The licensing side was more concrete: three years, more than 200 characters from Disney, Marvel, Pixar, and Star Wars, with Sora-generated fan videos on Disney+ tentatively scheduled for early 2026.

What actually shipped was much smaller. Sora grossed $2.14 million from 11.7 million downloads across iOS and Google Play, according to Appfigures Intelligence data reported by Ars Technica. The app peaked at roughly 3.3 million downloads in November 2025 and fell to 1.1 million by February 2026. For context, OpenAI closed a $110 billion funding round at a $730 billion valuation in February 2026 — the same month Sora hit its low point. The February round was led by Amazon, Nvidia, and SoftBank, all of whom have stronger interests in enterprise AI than in consumer video products.

The revenue gap was not a rounding error — it was a structural problem. Sora was burning through roughly $1 million every day in compute costs, not because the product had found an efficient distribution model but because video generation is intrinsically expensive to run at scale. Sam Altman made the call to shut Sora down — kill the consumer product, free up the GPU cluster, and refocus on coding tools and enterprise. Some OpenAI staffers on the Sora team were reportedly surprised when informed Tuesday morning. The Sora model research team itself is not being disbanded; the group is shifting toward robotics and physical world simulation, a move that aligns more directly with OpenAI's stated AGI mission and — not incidentally — has better unit economics.

then-CEO Bob Iger had signed the deal with Altman in December 2025, and by January 2026 was publicly discussing Sora-generated content appearing on Disney+. By the time OpenAI announced the shutdown on March 24, Disney had ended the partnership. Disney had not yet made any payments to OpenAI and no deal had been finalized — the $1 billion was a number in a term sheet, not money in an account. The company learned of the shutdown less than an hour before the public announcement, according to TechCrunch. A Disney spokesperson's statement to the Hollywood Reporter was diplomatic: "We respect OpenAI's decision to exit the video generation business and to shift its priorities elsewhere".

The OpenAI blog's original announcement of the deal made the conditions explicit: the transaction was subject to negotiation of definitive agreements, required board approvals, and customary closing conditions — none of which were completed. This was not a done deal that Disney walked from. It was a headline that got ahead of the paperwork.

The deal's formal structure — equity stake, character licensing, warrants to purchase additional shares — was confirmed in OpenAI's own disclosure, which described Disney taking a $1 billion equity stake alongside the licensing arrangement. The structure was unusual: a media company taking a large equity position in an AI lab while simultaneously licensing characters to that same lab's consumer product. That kind of multi-vector deal requires board approval from both sides, regulatory review in certain jurisdictions, and weeks of legal drafting. None of it happened before the plug was pulled.

For founders and investors, the lesson is not that AI video failed as a technology. Video generation is improving rapidly and is already embedded in creative workflows across film, advertising, and gaming. The lesson is that "strategic interest" and "signed contract" are different things, and that the gap between a partnership announcement and a revenue line can be the entire distance between ambition and execution. Disney walked from a deal without writing a check. The $1 billion was a headline, not an obligation. OpenAI will keep building — the compute freed by killing Sora will flow somewhere else, probably into agent products where the revenue-per-GPU math is more forgiving. The question founders should be asking is not whether the AI infrastructure buildout is real. It plainly is. The question is which consumer-facing products have the retention and monetization to survive the next round of reallocation.