California is the engine of global AI. It’s the ecosystem where the companies and technologies driving today’s AI revolution were born. But an environmental law passed in the 1970s for a radically different industrial economy, the California Environmental Quality Act (CEQA), is now threatening to stall that engine.

By subjecting virtually every large data center project, as well as the critical electric grid infrastructure needed to power them, to lengthy environmental reviews, public hearings, and extensive exposure to litigation, CEQA is creating regulatory uncertainty and delays at the heart of AI: data centers.

Artificial intelligence has become one of the most powerful tools of human progress, even as its broader impact is the subject of fierce debate. In response to that uncertainty, many policymakers have adopted a containment mindset, treating AI as something to be tightly regulated. That regulatory impulse, however, has increasingly shifted away from software and toward the infrastructure which makes AI possible, namely data centers and the energy that powers them.

When CEQA was created, it was designed for an economy in which electricity demand grew gradually, the grid had excess capacity, and environmental reviews were relatively predictable and fast. At the time, even large-scale projects like data centers could have obtained approvals and connected to the grid straightforwardly, without becoming trapped in endless litigation. Today, by contrast, the explosive growth in AI’s energy demand and the increasing legal weaponization of the permitting process have turned that same legal framework into a structural obstacle.

For decades, California benefited from relatively cheap energy, a grid capable of scaling with demand, and a regulatory environment which allowed new data centers to connect quickly to existing infrastructure. That combination made possible the concentration of computing power which underpinned California’s leadership in AI. But as many began using CEQA to challenge substations, transmission lines, and grid interconnections, what were once technical decisions gradually became legal battlegrounds. Each new project could now be delayed or blocked by lawsuits, favoring incumbents and pushing new investment out of the state.

In 2025, Governor Newsom vetoed AB 93, a bill requiring data centers to report water and energy use, warning it would drive away jobs and investment. CEQA, however, continued to impose the same structural constraints on new infrastructure.

More recently, in the Imperial Valley, the city of Imperial sued Imperial County to halt a major data center project, alleging CEQA violations and asking a court to suspend permits which had already been granted. The project’s developer then filed a federal lawsuit against the city, accusing it of administrative obstruction, leaving the investment tied up in court and highlighting the uncertainty which now surrounds this type of infrastructure.

Even large technology platforms cannot be taken for granted in such a volatile environment. Still, they are the ones best equipped to navigate it. Companies like Google, Amazon, and Microsoft can absorb these costs thanks to their large legal teams, deep capital reserves, and political influence. Startups and independent operators do not have the same luxury. They lack the resources to endure years of permitting delays and litigation, which effectively locks them out of the market and weakens competition in AI.

Under these conditions, California risks losing investment to states with more predictable regulatory regimes, especially Texas. By operating a competitive electricity market, allowing private developers to finance and build substations and transmission lines, and maintaining faster and more legally stable permitting processes, Texas has become one of the leading hubs for new data center development, attracting capital and AI related projects while California falls behind.