When Fidji Simo joined OpenAI as president and chief operating officer in mid-2025, the appointment was meant to signal maturity. Here was a seasoned operator—someone who had run Instacart through its post-pandemic recalibration and before that helped build Facebook's app ecosystem—arriving to professionalize the most consequential AI company on Earth. Her departure, confirmed this week with minimal fanfare, tells us that professionalization and OpenAI remain uneasy bedfellows.
The official framing is anodyne: Simo is leaving to pursue other opportunities, grateful for the experience, excited about the future. The subtext is considerably more interesting. OpenAI has now cycled through multiple senior executives in roles meant to bridge the gap between Sam Altman's product vision and the unglamorous work of actually running a company that employs thousands of people, burns billions of dollars, and operates under regulatory scrutiny on three continents.
The operator problem
OpenAI's challenge has never been technical talent. It has been organizational coherence. The company emerged from a nonprofit research lab, briefly became a capped-profit hybrid, and now operates as something closer to a conventional technology giant—except that its board dynamics nearly destroyed it in late 2023, its relationship with its largest investor remains perpetually complicated, and its competitive moat depends on staying ahead of well-resourced rivals who have access to the same fundamental research.
Simo was supposed to be the adult supervision that made all of this work. Her departure suggests either that the role was impossible, that the cultural fit never materialized, or that Altman's leadership style leaves little oxygen for a genuine number two. Probably some combination of all three.
What the pattern reveals
This is not an isolated incident. OpenAI has seen significant executive departures throughout its commercial era, from co-founders to safety researchers to operations leaders. Each exit has its own specific circumstances, but the pattern suggests a structural tension: OpenAI wants to be both a research organization pushing the boundaries of artificial intelligence and a commercial entity capable of generating the revenue to fund that research. These goals are not inherently contradictory, but they require different management approaches, different incentive structures, and different cultures.
Simo's background was pure commercial operator. She knew how to run logistics, manage headcount, optimize unit economics. What she apparently could not do—or was not permitted to do—was reshape OpenAI into a company where those skills could fully apply.
Our take
The departure matters less for what it says about Simo than for what it reveals about OpenAI's ongoing identity crisis. The company remains extraordinarily valuable, technically formidable, and culturally chaotic. It can launch GPT-5.6 and reshape entire industries while simultaneously struggling to retain the executives meant to manage that growth. For investors and partners, this should prompt uncomfortable questions: not about whether OpenAI can build impressive models, but about whether it can build an impressive organization. The former is proven. The latter remains very much in doubt.
