Eric Schmidt thinks the real limit to AI isn't energy but rather it's cash.
Milk Road AI Twitter · Milk Road AI (@MilkRoadAI) · 2026-05-24
Eric Schmidt argues that capital — not energy — is the true binding constraint on AI scaling, estimating approximately $50 billion is required per gigawatt of AI compute infrastructure.
Appears in
Extraction
Topics: ai-infrastructureai-investmentcompute-scaling
Claims
- Eric Schmidt argues that capital availability, not energy supply, is the primary bottleneck for AI development.
- Schmidt estimates approximately $50 billion is required per gigawatt of AI compute infrastructure.
- The widely held narrative that energy constraints are the binding limit on AI growth is incorrect, according to Schmidt.
Key quotes
Eric Schmidt thinks the real limit to AI isn't energy but rather it's cash.
His math is that roughly $50 billion per gigawatt of AI compute.