Add as a preferred source on GoogleAltman Repeats $20 Billion Goal for Five-Year AI Capacity Costs
Podcast host John Collison introduced $70 billion as a hypothetical comparison. Altman pointed to energy costs, supply and supply-chain margins as factors that would need to change.
OpenAI CEO Sam Altman said Thursday he would prefer the five-year cost of installing and operating one gigawatt of AI capacity to reach $20 billion, revisiting an ambition he has discussed before.
During a podcast discussion, Altman used $70 billion as a hypothetical five-year cost for one gigawatt, and host John Collison clarified that it covered installation and operation. Altman agreed with the framing and said he would be happier at $20 billion: “I think if that number were 20 instead of 70, we should be much happier.”
Altman said lower energy costs, increased supply and narrower gross margins across the supply chain would be needed to bring the total down. He also said the industry should continue pursuing 10-fold annual performance gains from smarter models.
Altman discussed a similar $20 billion goal in October 2025. His latest comments did not announce a new budget or construction plan. He gave no breakdown of the five-year cost or timeline for reaching it.
Simon Yoon
ReporterSimon Yoon reports on blockchain technology for TokenPost. Send corrections or tips to info@tokenpost.com.