Altman Targets AGI by Year-End as OpenAI Advances Agents and Hardware

OpenAI has evolved from an AI research laboratory into a consumer-product company, making access to chips and electricity central to its ability to train models and serve users. Chief Executive Sam Altman frames the industry’s product cycle in three waves: chatbots, coding agents and persistent agents that remain active like a chief of staff or colleague. The shift matters because those always-on systems would require far more inference capacity, tying OpenAI’s path toward artificial general intelligence, or AGI, to an infrastructure build-out.

In an interview with Ti Morse on the roughly 70-minute Relentless podcast, covered on July 26, 2026, Altman said he had “seriously underbet” on computing capacity after being intimidated by financial markets. He identified chips as OpenAI’s immediate constraint and electricity as the next. A related estimate cited in the report put OpenAI’s compute spending through 2030 at as much as $750 billion, against revenue of about $25 billion. Altman disclosed no new investment figure but said persistent agents could arrive soon and argued AGI would make people busier, not idle.