Barret Zoph, who returned to OpenAI in January after a brief stint at a competing startup, is leaving the company again after only five months, according to internal communications and company confirmations. His departure marks yet another high-profile exit from the AI giant, which has seen significant leadership turnover in recent years.

Zoph rejoined OpenAI in mid-January as head of enterprise AI sales, a role that placed him at the center of the company's push to generate revenue from business clients. He had previously left OpenAI in the fall of 2024 to become co-founder and chief technology officer of Thinking Machines Lab, a rival firm founded by former OpenAI chief technology officer Mira Murati. That venture itself was born from a wave of departures that followed Murati's exit from OpenAI in September 2024.

Background on Barret Zoph

Before his initial departure from OpenAI, Zoph was a key technical leader at the company. He joined OpenAI in 2023 after several years at DeepMind, where he contributed to foundational research in reinforcement learning and natural language processing. At OpenAI, he worked on alignment and safety projects before moving into product-focused roles. His known contributions include work on early versions of GPT-4 and the development of reinforcement learning from human feedback systems.

Zoph's career in artificial intelligence began at the University of California, Berkeley, where he earned a PhD in computer science. His doctoral research focused on neural architecture search and automated machine learning, areas that later proved crucial to scaling large language models. After graduation, he joined DeepMind, where he was part of the team that developed AlphaFold and other landmark models.

The Thinking Machines Lab Interlude

On leaving OpenAI in 2024, Zoph became part of a major talent exodus that included several senior researchers and engineers. Murati’s Thinking Machines Lab was formed with the stated goal of building safe, advanced AI systems that could be deployed in high-stakes environments. Zoph served as CTO from October 2024 until January 2026, when reports emerged of alleged misconduct involving an undisclosed relationship with a colleague. Murati posted on the social platform X in January that the company had “parted ways” with Zoph and that he would be replaced as CTO.

The circumstances surrounding his departure from Thinking Machines Lab remain unclear, as neither Zoph nor the company have provided further details. However, the allegations prompted a rapid return to OpenAI, where he was welcomed back alongside two other former Thinking Machines Lab employees: Luke Metz and Sam Schoenholz. OpenAI’s CEO of Applications, Fidji Simo, wrote on X at the time that the decision had “been in the works for several weeks.”

Return and Renaissance at OpenAI

When Zoph returned to OpenAI in January 2026, the company was undergoing a strategic refocus. Like many AI companies, OpenAI had been racing to capture market share in enterprise AI, competing with Google, Amazon, and a host of startups. Under a new directive from CEO Sam Altman, the company vowed to stop chasing what he called “side quests” and instead double down on revenue-generating products: enterprise AI tools, coding assistants, and other commercial services.

Zoph’s role as head of enterprise AI sales was central to this vision. His job was to lead the sales force that would convince large businesses to adopt OpenAI’s platform for tasks like customer support automation, data analysis, and workflow optimization. The enterprise market represents a massive opportunity; industry analysts project it will be worth tens of billions of dollars by 2030.

Despite this promising mandate, Zoph’s second tenure at OpenAI was short-lived. Company insiders say he posted a farewell message in internal Slack channels over the weekend, thanking colleagues and expressing confidence in the team’s future. OpenAI confirmed his departure to The Verge but declined to provide further detail about the reasons behind it. Zoph did not respond to interview requests.

Broader Context of OpenAI Turmoil

Zoph’s exit is the latest in a series of leadership changes that have affected OpenAI since the dramatic boardroom fight in November 2023. During that period, CEO Sam Altman was briefly ousted by the board, leading to a wave of resignations and threats from employees. Altman was reinstated within days, but the incident exposed deep fractures within the company. Murati briefly served as interim CEO during the upheaval, and she later testified in a legal proceeding that she could not trust everything Altman said.

The fallout continued through 2024 and 2025. Several senior figures left to start their own ventures or join competitors. In addition to Murati, co-founder Ilya Sutskever departed to form Safe Superintelligence Inc. Other notable exits included research leads and policy experts.

OpenAI’s ability to retain top talent is critical as it prepares for an initial public offering, which sources say could occur as early as 2027. The company is reportedly seeking a valuation exceeding $300 billion, making it one of the most valuable private companies in the world. Investors are closely watching employee turnover, as high departures can signal instability and affect long-term growth prospects.

Enterprise AI and the Road Ahead

The enterprise segment is particularly important for OpenAI’s IPO narrative. By selling AI services to businesses, the company can demonstrate a predictable revenue stream that is less dependent on consumer subscriptions. OpenAI has already secured key partnerships with major banks, retailers, and tech corporations. Its enterprise platform includes tools for document processing, code generation, and custom model fine-tuning.

Competitors like Google DeepMind, Anthropic, and Microsoft are also aggressively targeting the same market. Microsoft, which is OpenAI’s largest investor, has integrated some of OpenAI’s models into its Azure cloud offerings but also develops its own AI products. This creates both a partnership and a rivalry that complicates OpenAI’s market positioning.

With Zoph gone, OpenAI will need to appoint a new head of enterprise sales quickly. The company has a deep bench of executives, but replacing a senior leader who has now left twice in two years could prove challenging. Some observers speculate that internal culture issues—stemming from fast growth, intense pressure, and a high-stakes environment—may be contributing to the churn.

The three returning employees—Zoph, Metz, and Schoenholz—were seen as a signal that OpenAI could still attract talent back even after the Thinking Machines Lab exodus. Now only two remain, and the company must work to retain them and other key staff if it hopes to maintain momentum.

What Might Come Next for Zoph

Barret Zoph’s career path is a case study in how the AI talent wars are playing out. He has worked at two of the most prestigious AI labs in the world, co-founded a startup, and returned to his previous employer—only to leave again. It remains to be seen whether he will start another company, join a different incumbent, or take a break from the industry. Given his expertise in both research and product strategy, he is likely to have multiple offers.

For now, the AI community is watching closely. Zoph is known for his quiet, technical demeanor and was never a public face of OpenAI. But his roles at the company, first in research and later in enterprise, highlight the breadth of skills that top AI talent must possess. His departure may accelerate conversations within OpenAI about how to create a more stable and supportive work environment, especially for employees who rejoin after time away.

In the competitive landscape of AI, stability is a luxury few companies can afford. OpenAI will need to navigate these challenges if it wants to keep its position as the leader in generative AI.

Source: The Verge News