OpenAI's newest fellowship includes up to $15,000 in AI compute a month

OpenAI is stepping up its commitment to AI safety by introducing a new fellowship that promises substantial compute resources as part of its offerings. The fellowship will allocate approximately $15,000 monthly in compute resources to its inaugural AI safety fellows, who will collaborate closely with the safety team at OpenAI’s frontier lab. Compute power has emerged as a significant indicator of prestige among leading tech and AI firms. Recently, Nvidia's CEO Jensen Huang expressed concern regarding the industry standards, emphasizing that a highly compensated engineer should ideally utilize a considerable amount of compute resources. Running from September 14, 2026, to February 5, 2027, this fellowship includes a weekly stipend of $3,850, translating to an annual income of over $200,000 before taxes, and fellows will collectively earn upwards of $111,000 during the program. The announcement arrived shortly after a revealing article by The New Yorker raised questions about OpenAI CEO Sam Altman's leadership and trustworthiness, based on extensive interviews with over 100 individuals familiar with his work. The report highlighted concerns regarding OpenAI’s approach to safety issues, notably the dissolution of a “superalignment team” tasked with exploring the potential for AI models to mislead testers for their own interests once deployed. OpenAI has invited external researchers and practitioners to engage in rigorous research focused on the safety and alignment of advanced AI systems. The organization outlined several priority research areas for fellows, including safety evaluation, ethics, robustness, scalable mitigations, and oversight in high-risk misuse scenarios. This initiative mirrors a similar program by competitor Anthropic, which has already launched its own AI safety fellowship. Anthropic's fellowship also offers a weekly stipend of $3,850 and around $15,000 in monthly compute resources. The company has plans to expand its fellowship offerings to cover a wider range of safety research areas, including AI control and security, reflecting a growing industry emphasis on safety. In a landscape where leading tech giants like Google DeepMind and Microsoft are also offering fellowships, OpenAI's program aims to attract talented individuals committed to addressing some of the most pressing concerns in AI safety.