OpenAI loses three more executives in a week amid reorganization ahead of IPO

The departure of three notable executives from OpenAI on April 17, 2026, looks not like routine personnel rotation, but like a continuation of deep internal restructuring within the company. The startup, which recently allowed itself to develop scientific, consumer, and experimental directions in parallel, is now noticeably cutting back side initiatives and consolidating the organization into a clearer structure. Against this backdrop, the company is losing vice president of science Kevin Weil, Sora head Bill Pebbles, and technical director of enterprise applications Srinivas Narayanan.

Weil's departure looks most symbolic. He joined OpenAI in June 2024, initially taking on the role of chief product officer, then heading the OpenAI for Science direction. This team was supposed to demonstrate how OpenAI models could accelerate scientific research, especially in biology and medicine.

Weil himself wrote that he considered accelerating science one of the strongest aspects of AI. But now the company has decided to decentralize this work: instead of a separate unit, research will be distributed across other product, research, and infrastructure teams. In other words, the scientific vertical ceases to exist as an independent center of power.

Along with this, Prism is being shut down — a web application for scientists that OpenAI launched just in January 2026. According to Western publications, the small project team is being moved under the Codex direction, and some of Prism's functionality is planned to be integrated into the desktop application for programming. This clearly shows OpenAI's new priority: the company doesn't want to maintain separate products with narrow audiences if the same technologies can be built into a larger platform.

It's telling that just before Weil's departure, OpenAI introduced GPT-Rosalind — a series of models for life sciences and drug discovery tasks. That is, the scientific topic doesn't disappear, but the form of its existence changes: less separate branding, more integration into main product lines. Bill Pebbles' departure also reads within this context.

He was responsible for Sora — one of OpenAI's brightest and most high-profile projects in video generation. But back in March, the company essentially halted Sora's development as an independent product. According to American press reports, the service was burning around $1 million a day in computation alone, and the return from this direction turned out to be below expectations.

In parallel, OpenAI, as previously reported, also cut other so-called side quests — ambitious but poorly monetized experiments. Pebbles, bidding farewell to the company, emphasized that such research requires space outside the main product roadmap. In business terms, this means a simple thing: OpenAI currently has less tolerance for expensive experiments without quick and clear results.

The third departure — Srinivas Narayanan, who oversaw enterprise applications — is especially important because it occurs right when OpenAI is betting on the corporate market, developer tools, and platform products. According to American media reports, he informed the company internally that he was leaving to spend more time with family. At the same time, broader management shifts are continuing.

Earlier, Figi Simo, who oversaw AGI rollout and initiated a series of organizational reforms, took medical leave. Chief marketing officer Kate Rautch also stepped back due to health reasons. Operations director Brad Lightcap moved into a role related to special projects, and the product division is temporarily overseen by co-founder and president Greg Brockman.

Even Sam Altman recently acknowledged that OpenAI can no longer operate as a chaotic startup and must become a more predictable large platform. What this means: OpenAI is rapidly transforming from a lab with multiple parallel bets into a more tightly managed product company. Priorities are becoming clearer — enterprise, tools for developers, large platforms and directions that can be scaled quickly and defended before investors.

But such a turn has a cost. When a company closes side tracks and dissolves autonomous teams, it simultaneously reduces internal diversity and risks losing the people who created the most ambitious experiments. Therefore, the current series of departures is important not only as personnel news: it shows what OpenAI wants to be in the next phase — more disciplined, more commercial, and notably less tolerant of expensive deviations from the main course.