OpenAI Aims to Deploy Legitimate AI Researcher by 2028, Says Sam Altman

OpenAI reports rapid progress in its deep learning systems, with models growing increasingly adept at solving complex problems at an accelerating pace. In fact, CEO Sam Altman stated during a Tuesday livestream that the company is on track internally to develop an intern-level research assistant by September 2026, followed by a fully automated, "legitimate AI researcher" by 2028.

This ambitious roadmap coincides with OpenAI's final transition to a public benefit corporation structure, marking a shift from its non-profit origins. The change frees the company from constraints associated with its non-profit charter and unlocks new avenues for raising capital.

Chief Scientist Jakub Pachocki, who joined Altman on the stream, described the envisioned AI researcher—distinct from a human AI researcher—as a "system capable of autonomously completing large-scale research projects."

"We believe deep learning systems could achieve superintelligence within the next decade," Pachocki added. He defined superintelligence as systems that surpass human capabilities across a broad spectrum of critical tasks.

To reach these objectives, OpenAI is focusing on two core strategies: ongoing algorithmic innovation and a massive increase in "test time compute"—the duration models spend reasoning through problems. Pachocki noted that current models can manage tasks with about a five-hour time horizon and rival top humans in competitions like the International Mathematical Olympiad. He anticipates this horizon will expand quickly, partly by allocating far greater computational resources for models to ponder complex challenges. For significant scientific discoveries, he suggested it would be worthwhile to dedicate an entire data center's computing power to a single problem.

OpenAI states these targets align with its broader mission to accelerate scientific research. The goal is for AI to potentially make discoveries faster than human researchers, address problems currently beyond human capability, and significantly hasten technological innovation in fields like medicine, physics, and tech development.

Altman added that the corporate restructuring establishes a framework to support this aggressive timeline for AI research assistants while upholding a commitment to responsible development. Under the new structure, the non-profit OpenAI Foundation, focused on scientific progress, will own 26% of the for-profit entity and govern its research direction. The foundation also has a $25 billion commitment to using AI for disease cures and will help oversee AI research and safety initiatives.

2-FOR-1 DISCOUNT: Bring a +1 and save 60%

Hear from over 250 industry leaders from Google Cloud, Netflix, Microsoft, Box, Phia, a16z, ElevenLabs, Wayve, Hugging Face, Elad Gil, Vinod Khosla, and more. They'll lead 200+ sessions packed with insights to fuel startup growth and sharpen your competitive edge. Plus, explore 300+ showcasing startups across all sectors.

Bring a colleague and save 60% on their pass, or secure your own pass by October 27 to save up to $444.

2-FOR-1 DISCOUNT: Bring a +1 and save 60%

Hear from over 250 industry leaders from Google Cloud, Netflix, Microsoft, Box, Phia, a16z, ElevenLabs, Wayve, Hugging Face, Elad Gil, Vinod Khosla, and more. They'll lead 200+ sessions packed with insights to fuel startup growth and sharpen your competitive edge. Plus, explore 300+ showcasing startups across all sectors. Bring a colleague and save 60% on their pass, or secure your own pass by October 27 to save up to $444.

According to Altman, the for-profit arm's enhanced fundraising capacity will enable it to scale up the necessary infrastructure to achieve these scientific breakthroughs. He revealed that OpenAI has committed to building 30 gigawatts of infrastructure—a financial undertaking of approximately $1.4 trillion—over the coming years.