California is staring down a major energy reckoning. According to a recent Stanford roundtable of experts, by 2040, the state’s peak power demand could rise by an amount equal to what’s needed to power 20 million more homes, more than the number of current homes in California, with AI data centers being one of the biggest drivers of that surge.
The escalation is not just about electricity. It comes with complications: permitting delays, slow grid connections, and regulatory friction are already hampering the build-out of data centers and associated infrastructure. Experts warn that unless California’s approval processes speed up and its policies become more aligned with the rapid pace of AI growth, the state risks losing investment and talent to other regions with more streamlined regimes.
During the meeting, stakeholders from utilities, tech, government, finance, and academia came together to map a path forward, laying out five strategic levers needed to keep the state competitive while meeting demand sustainably. One major recommendation is creating a one-stop permitting venue that merges local, state, and federal reviews, so that planning, approvals, and grid interconnection aren’t fragmented and time-consuming.
Others include speeding up grid connection processes for data centers and energy storage facilities, reducing delays for new power plants and transmission lines, and making investments that balance affordability, reliability, and decarbonization. The experts spelled it out clearly: California has roughly 24 months to implement changes or risk falling behind in the AI infrastructure race.
If the state acts decisively, the rapid increase in electricity demand could translate into economic growth. If not, it may face costly delays, energy unreliability, and a lost opportunity to lead in AI infrastructure. The stakes are high and so are the decisions ahead.