OpenAI CEO Sam Altman has unveiled an ambitious vision to massively scale artificial intelligence infrastructure, proposing what he calls a “gigawatt factory” capable of producing a gigawatt of new AI infrastructure every week.
In a blog post titled ‘Abundant Intelligence’, Altman highlighted the unprecedented growth in AI usage and predicted even greater demand in the future. He argued that access to AI could become as fundamental as access to electricity or the internet and possibly even recognized as a human right.
“Almost everyone will want more AI working on their behalf,” Altman wrote, stressing that both inference compute to run models and training compute to improve them must be scaled significantly to meet global needs.
Altman painted a picture of what could be possible if compute power expands dramatically: with 10 gigawatts of compute, he suggested, AI might help cure cancer or deliver personalized tutoring to every student worldwide. But without enough infrastructure, hard trade-offs would need to be made. “No one wants to make that choice, so let’s go build,” he said.
The proposed gigawatt factory would require innovation across every layer of the technology stack, from chips and power generation to robotics and construction. Altman acknowledged the scale of the challenge, saying it could take years to achieve the milestone, but called it “the coolest and most important infrastructure project ever.”
He emphasized the importance of building much of this infrastructure in the United States, noting that other countries are currently moving faster in areas like semiconductor fabs and energy production. OpenAI, he said, aims to help reverse that trend.
Over the coming months, Altman said OpenAI will share more details about its plans and partnerships to bring the project to life. Later this year, the company expects to announce financing strategies, hinting at “interesting new ideas,” given the direct link between compute capacity and revenue growth.
