OpenAI CEO Sam Altman has offered his perspective on why prominent researchers and executives leave established AI companies to launch competing ventures. In a podcast conversation reported by The Times of India on November 8, Altman suggested that people may feel they can only trust themselves to build a technology they see as unusually consequential.
The discussion took place on a podcast hosted by Raj Shamani. Altman was asked why the AI industry has seen friction and why people associated with OpenAI have gone on to build other ventures, even though the company helped popularise modern generative AI.
He connected departures to both disagreements over how AI should be built and concerns about who holds decision-making power. In his account, those issues can overlap: a person may believe a company is following the wrong approach and also worry about allowing one leader or organisation to control the direction.
The comments are Altman’s interpretation of a complicated set of individual decisions. They do not explain every departure from OpenAI, and the people who left may have had different reasons. Some may have wanted to pursue their own research priorities, build new products or create organisations with different structures.
The AI industry has produced a number of spin-offs and new companies founded by researchers and executives with experience at established labs. Those founders bring expertise and networks, but they also face the challenge of competing for computing resources, talent, data and customers.
Altman’s comments raise a broader governance question: how should powerful AI systems be developed, and who should have influence over their design and deployment? Companies need decision-making structures that can set priorities, respond to risks and deliver products. Researchers may also want room to explore approaches that do not align immediately with commercial timelines.
The balance can be difficult. A company may prioritise improvements to models already in use, while researchers may want to test alternative methods or investigate longer-term safety issues. Those priorities are not necessarily incompatible, but they can compete for budget, computing capacity and leadership attention.
OpenAI itself has experienced a highly visible leadership crisis. Altman was removed as CEO by the board in November 2023 and later reinstated. The Times of India report noted that several prominent figures subsequently left the organisation, including co-founders and senior leaders. The context has contributed to public discussion about trust and corporate governance in AI development.
However, leadership departures do not automatically indicate that a company’s research is failing or that a new startup will succeed. New organisations must demonstrate that their technology works, attract investment and build a sustainable business. Established firms retain advantages in infrastructure, deployment and access to customers.
For users and policymakers, the concentration of AI capability in a small number of organisations raises questions about accountability and public oversight. Decisions made by companies can affect how systems are trained, what safeguards are applied and which applications are released.
Altman’s remarks offer one explanation from the perspective of an executive leading a major AI company. They do not settle the debate over whether AI development should be concentrated or decentralised. The continuing movement of researchers between large labs and startups shows that the industry is still searching for organisational models that can support both innovation and responsible control.