infrastructureScore 45/100Research
Data center buildout critical infrastructure bottleneck; massive capex acceleration required for AI training and inference separation
Leopold Aschenbrenner· Forethought· AI· 2026-07-21
“So they retrofit the existing data centers to serve inference. People can still keep talking to their AI agents, but they're going to stop getting better and better for like 6 months to a year while they build the new data centers that are going to be the transparent data centers... this whole time they're building more and more data centers more and more chips”
Why it matters
Aschenbrenner explicitly identifies data center construction as a critical bottleneck in AI development. His regulatory scenario requires dual data center infrastructure (inference vs. training), doubling capital requirements.
Investment implication
Massive capex acceleration for data center operators, power infrastructure, cooling systems, and semiconductor manufacturing. Companies supplying colocation, power delivery, and thermal management face sustained demand growth through 2030s.