IEA: Energy and AI
IEA's base case projects global data center electricity consumption around 945 TWh by 2030, growing much faster than other electricity demand.
Sources
Market evidence behind the SOLAI infrastructure thesis.
The site uses third-party research to frame why AI compute decentralization matters: capital intensity, power constraints, inference growth and infrastructure concentration.
McKinsey: $7T Data Center Build-Out
McKinsey estimates global spending on data centers could reach $7 trillion by 2030, driven heavily by AI infrastructure requirements.
Bain: AI Compute Requirements
Bain reports that global incremental AI compute requirements could reach 200 GW by 2030 and that $2 trillion in annual revenue may be needed to fund scaling.
Goldman Sachs: Power Demand
Goldman Sachs Research projects data center power demand to surge 175% by 2030 versus 2023 levels, highlighting power, policy and supply constraints.
Why these numbers matter
Centralized AI infrastructure is becoming a capital, energy and availability bottleneck. SOLAI's thesis is that useful AI capacity should not only come from hyperscale data centers. Local machines, independent providers and agent-managed temporary clusters can become part of a broader compute fabric.