Naveen Rao: 4D Computing, AI's Energy Wall & Beating Biology

Naveen Rao says AI's binding constraint is energy, not GPUs, and unveils a prototype "dynamical computer" targeting 1,000x power efficiency in 3.5 years.

• Energy is the real bottleneck: Naveen Rao says Google alone processes 3.2 quadrillion tokens/month, implying ~12 gigawatts of AI energy demand versus ~40 gigawatts total US data center capacity.
• 50% of token cost is power: Rao states energy now accounts for half the cost of serving a ChatGPT-style token, with the rest being CapEx on hardware and floor space.

Themes, sentiment and reasons are AI-written paraphrases, checked so they never quote an episode. They may contain errors. Not investment advice. Episodes remain the property of their creators.