Why compute might get 10x+ more expensive in coming years

If a human-level software engineer that could run on an H100 equivalent, at current market rates for software engineers, that H100 should rent for over $250k a year. That’s 15x today’s spot price.

Open original article View publication

What's interesting

Engagement

Likes: 533

Comments: 92

Restacks: 55

Entities

Topics

Moore's Law EUV GPU AI compute economics Anthropic revenue growth Compute scaling Inference margins Economies of scale in AI Compute pricing Moore's Law and fab capacity AI infrastructure investment

People

Paul Ehrlich

Companies

OpenAI GB300 GB200 H100 SpaceX Google Anthropic

Sponsorship

No sponsorship detected

8 Predictions for the Era of Continual Learning

Dwarkesh Podcast · Google · Same publication

Dylan Patel – Anthropic & OpenAI will have most of the world’s compute by 2028

Dwarkesh Podcast · Google · Same publication

🍻 Tretas da bolha tech da semana

@manodeyvin · Google

The AI Trade is a Psyop

Notes from the Circus · Google

The Tribute

Notes from the Circus · Google

When Biden Dies

Political Currents by Ross Barkan · Google

Meta Settled a Massive Lawsuit Around Harming Kids. Then It Threw Itself a Parade.

Techish · Google

Your creator is on competitors radar

The Playbook · Google