What Should a Gigawatt of AI Cost?

Sam Altman wants dollars per gigawatt of leading-edge AI to fall, and he wants the five-year cost of one gigawatt to drop from $70 billion to $20 billion. The cost of turning sand into intelligence should fall with it: energy, chips, and the supply chain should all get cheaper. If the cost does not fall, he will be very sad. Supply has to rise sharply because there is so much value in AI.

People already give different five-year ownership costs for a gigawatt. He picks $70 billion as the cost to install one and run it for five years. At 20 instead of 70, everyone should be much happier, and models should still deliver 10x better performance every year. He does not think a cost of 2 will happen. Even $20 billion means energy has to get cheaper, and a lot of the gross margin between melting sand and the token has to get compressed. He does not know the math off the top of his head. If all materials were free and no one made a profit, he guesses the current energy envelope would only just cover the cost.

He wants a compute glut: far more memory, far more data centers, and far more of every input, so anyone who wants a computer can have one. A rising price per gigawatt would mean compute is not keeping up. He believes deflationary technology is good.