OpenAI killed a billion-dollar Sora partnership with Disney less than an hour before the public announcement, according to a Wall Street Journal investigation cited by TechCrunch. That is the fact that makes the company's "focus era" narrative fall apart.
Kevin Weil, who as of today is no longer at OpenAI, spent seven months running OpenAI for Science. He launched it last September as a flagship initiative to accelerate scientific discovery, one of what he called the clearest ways AI could benefit humanity. As of today, it is being dissolved into other research teams. The man who built it is gone. The project is dead.
Weil is not the only exit. Bill Peebles, who built Sora from zero to launch, also left today. OpenAI discontinued Sora on March 24, six weeks after launch, when the app was burning through roughly $1 million in compute costs per day and user count had collapsed from a peak around 1 million to under 500,000. Disney had committed $1 billion to a Sora partnership. The entertainment giant found out the deal was dead less than 60 minutes before the public did, the WSJ found.
The pattern is not subtle. OpenAI launched Prism, a web app for scientists, in January. It is being discontinued. The roughly 10-person team behind it is being folded under Thibault Sottiaux, head of Codex, which OpenAI intends to build into an "everything app" for the enterprise market. OpenAI confirmed it is unifying its business around enterprise offerings and coding, a pivot it describes as focus.
The departure of Weil is particularly clarifying. He did not join OpenAI to ship enterprise software. He joined as chief product officer, then moved to the research team to start OpenAI for Science, then left as that initiative was decentralized into other teams. His exit is the third high-profile departure of a senior executive in recent weeks, alongside Srinivas Narayanan, OpenAI's chief technology officer of enterprise applications, who is also leaving.
OpenAI has framed the restructuring as a deliberate consolidation. The numbers tell a different story. Sora cost $1 million every day to run. Prism launched in January and is dead by April. OpenAI for Science launched in September and is dead by April. These are not the timelines of a company in control of its product roadmap. These are the timelines of a company catching up to its own spending.
Sam Altman acknowledged as much in a recent blog post. "I am also very aware that OpenAI is now a major platform, not a scrappy startup, and we need to operate in a more predictable way now," he wrote. He did not mention Sora by name. He did not mention Weil. The post was about everything except the exits that were already in motion.
OpenAI is preparing for an IPO at a $852 billion valuation, a figure some investors are already questioning. The company raised $122 billion in its most recent funding round and says that shows strong conviction in its direction. Whether the enterprise pivot produces sustainable revenue is the only question that matters now. The side quests are gone. The question is what OpenAI actually built in their place.
