Barret Zoph has left OpenAI after five months in charge of enterprise AI sales, just as the company is trying to prove it can turn ChatGPT into a durable business product, not only a consumer habit.
Barret Zoph is out at OpenAI, again. The Verge reported on June 19, 2026 that Zoph has departed after returning in mid-January to lead the company’s enterprise AI sales push, a role OpenAI had positioned around business customers and coding products ahead of a planned IPO. OpenAI confirmed the departure to The Verge. Zoph posted a goodbye message internally and did not immediately comment publicly.
You don’t need to be inside OpenAI to see why the timing is poor. The company has been telling the market that enterprise customers are central to its next phase, while its consumer products, research ambitions and infrastructure bills all compete for attention. Axios reported in March that OpenAI said it was on pace for $25 billion in revenue this year, but also noted Ramp data showing Anthropic taking more than 73% of AI tool spending among companies buying AI tools for the first time. That is the customer OpenAI wants. Losing the executive put in charge of winning that customer after five months is not a small personnel note.
Zoph’s return was already complicated. He had left OpenAI in 2024 for Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati, then came back to OpenAI in January 2026 with Luke Metz and Sam Schoenholz. The Wall Street Journal reported at the time that Zoph had been fired from Thinking Machines after an undisclosed workplace relationship and after what Murati described internally as performance, conduct and trust issues. Zoph’s side was different: he said he was pushed out after Murati learned he planned to leave. Those two accounts still sit there, unresolved in public.
OpenAI hired him back anyway. Frankly, that is the part worth sitting with. If the concerns from Thinking Machines were serious, OpenAI accepted a reputational risk in a customer-facing role. If they were not serious, OpenAI still installed a leader carrying a public dispute into one of the most commercially sensitive jobs in the company. Either way, the result is the same: the person brought in to give enterprise sales more force is gone before he has had time to make the job his own.
The Thinking Machines story also gives this exit a sharper edge. Reuters reported in July 2025 that Murati’s startup raised $2 billion at a $12 billion valuation, with backers including Nvidia, AMD, Cisco and Jane Street. That was not a tiny lab playing with a side project. It was supposed to be one of the main post-OpenAI power centers, built around people who knew how frontier AI companies actually work. Instead, its early story has been dominated by departures, including Zoph and Metz back to OpenAI, Schoenholz also returning, and Andrew Tulloch leaving for Meta.
You can call that normal movement in a hot market, but only up to a point. Senior AI talent moves because the money is large, the equity swings are enormous, and the best researchers can change a company’s technical direction. Still, constant motion has a cost. Customers buying enterprise AI do not only buy a model score on a benchmark. They buy a roadmap, a support structure, security reviews, procurement patience and a person inside the vendor who will still be there when the contract renewal comes around.
That is where Zoph’s exit hurts OpenAI most. The company can absorb almost any single departure on paper. The Guardian reported in March that OpenAI closed a $122 billion funding round at an $852 billion valuation, and the company still has Sam Altman, Greg Brockman, Fidji Simo, Denise Dresser and a brand that most enterprise buyers already know. OpenAI is not short of money or attention.
But enterprise buyers are not impressed by drama for long. They may tolerate it when the product is far ahead of everyone else, yet the gap is no longer as simple as it looked in 2023. Anthropic has become a serious enterprise rival. Google has Gemini inside Workspace. Microsoft still wraps Copilot around the software many companies already use every day. OpenAI has to sell stability now, not just magic.
That makes the next appointment important. OpenAI does not need another famous research name cycling through a commercial job for a few months. It needs someone who understands procurement, data security, account expansion and the slow work of making a large customer feel looked after. Salesforce did not become Salesforce because buyers liked changing account leadership every quarter.
Zoph’s exit will not derail OpenAI by itself. Nothing in this story suggests that. But it does expose a real tension inside the company’s commercial turn: you cannot ask enterprises to bet more of their operations on your platform while the leadership around that platform keeps changing. The models may improve every few months. The people selling them cannot operate on the same release cycle.
Also read: Geoffrey Hinton says today's chatbots are already conscious and the AI safety debate may never be the same, RMZ Group is staking $35 billion on becoming the landlord of India's AI era, Yann LeCun calls xAI a failure as Elon Musk's lab pivots from frontier AI to renting out servers
