OpenAI Reveals Next Flagship Model Codenamed "Spud" — Altman Hands Off Safety to Focus on Infrastructure
OpenAI CEO Sam Altman told staff this week that the company has wrapped initial development on its next flagship model, internally codenamed "Spud" — a system Altman described as capable of "accelerating the economy." The announcement came the same day OpenAI confirmed it is shutting down Sora, the AI video generator it launched with significant fanfare just over a year ago, as the company redirects resources toward its next model generation and an anticipated IPO.
The structural news may ultimately matter more than the codename. Altman simultaneously stepped back from direct oversight of OpenAI's safety and security teams, handing that responsibility to VP of Research Mark Chen under the research division. It's the most consequential change to OpenAI's safety governance since co-founder Ilya Sutskever departed, and it shifts safety from a standalone function to one embedded within the research org — the exact configuration critics have long warned would subordinate safety accountability to product velocity in a competitive, capital-intensive environment.
Altman's stated reason for the move is straightforward: he needs to focus on raising capital, managing supply chains, and building what he described as data centers "at unprecedented scale." With an IPO on the horizon and a next-generation model to ship, the tradeoffs OpenAI is making — and who is accountable for them — are now a matter of public record in a way they weren't before.