OpenAI commits $1.4T to build AI researchers by 2028

OpenAI CEO Sam Altman announced that the company expects to develop an “intern-level research assistant” by September 2026 and a fully automated “legitimate AI researcher” by 2028. The ambitious timeline coincides with OpenAI’s transition to a public benefit corporation structure, which removes non-profit limitations and opens new capital-raising opportunities to fund the massive infrastructure needed for these AI breakthroughs.

What you should know: OpenAI is tracking toward creating AI systems capable of conducting independent research projects within the next few years.

• Chief scientist Jakub Pachocki described the AI researcher as a “system capable of autonomously delivering on larger research projects,” distinct from human AI researchers.
• Current models can handle tasks with roughly five-hour time horizons and match top human performers in competitions like the International Mathematics Olympiad.
• The company believes deep learning systems could achieve superintelligence—systems smarter than humans across critical actions—within a decade.

The big picture: OpenAI’s restructuring creates a framework to support aggressive AI development while maintaining research oversight.

• The non-profit OpenAI Foundation will retain 26% ownership of the for-profit arm and govern research direction.
• The foundation has committed $25 billion specifically for using AI to cure diseases and will manage AI research and safety initiatives.
• This structure allows the for-profit division to raise funds necessary for scaling infrastructure to achieve scientific breakthroughs.

Key technical strategy: OpenAI is betting on two main approaches to reach these milestones: continued algorithmic innovation and dramatically scaling “test time compute.”

• Test time compute refers to how long models spend thinking about problems before providing answers.
• For major scientific breakthroughs, Pachocki said it would be worth dedicating entire data centers’ worth of computing power to single problems.
• The extended thinking time is expected to rapidly expand the time horizon models can handle for complex tasks.

The infrastructure commitment: Altman revealed OpenAI has committed to building 30 gigawatts of infrastructure over the next few years.

• This represents a $1.4 trillion financial obligation that the new corporate structure is designed to support.
• The massive scale reflects OpenAI’s belief that achieving AI research capabilities requires unprecedented computational resources.

Why this matters: These AI research systems could potentially make scientific discoveries faster than human researchers and tackle problems beyond current human capabilities, dramatically accelerating technological innovation across fields like medicine, physics, and technology development.