California’s biggest utility provider, PG&E, has reported a dramatic rise in power supply inquiries from data center developers—marking an over 40% increase in demand this year alone. According to company executive Mike Medeiros, this uptick is concentrated across Northern California, including the tech-heavy Silicon Valley. Despite concerns that high real estate prices and electricity costs might deter new builds, especially for AI-focused data centers, PG&E is seeing renewed momentum from developers aiming to expand operations within the state.

Although analysts have long argued that California's expensive landscape limits its competitiveness against more affordable regions like the central U.S., PG&E insists there is still considerable opportunity left in the Golden State. The utility, which serves around 16 million customers across northern and central California, recently initiated a new application round for developers hoping to connect data centers to its infrastructure. The result was a robust response, with an additional 4.1 gigawatts of interest on top of the previously reported 8.7 gigawatts, signaling a clear upward trend in activity.

Moreover, not only is the number of potential projects rising, but their scale is significantly increasing as well. Last year, the average data center proposal ranged between 50 to 100 megawatts, but the latest proposals now stretch to 500 megawatts and even 1,000 megawatts in capacity. Medeiros attributed this shift to both the availability of land and the growing efficiencies tied to building large-scale facilities—a particularly critical factor for AI applications, which are notoriously power-hungry.

Unlike traditional data centers located near urban hubs for consumer-facing cloud services, AI-oriented facilities are primarily focused on training large language models such as ChatGPT. These operations require immense energy loads and benefit from being situated in more remote or less densely populated areas. As a result, many of the proposed new builds are not located in Silicon Valley but are instead shifting inland, with Contra Costa County and Fresno emerging as hot spots for upcoming development.

Despite the surge in interest, PG&E remains realistic that not all inquiries will result in final connections. One of the largest obstacles for the utility is meeting the tight construction and commissioning timelines set by tech firms. Additionally, there are supply chain constraints—particularly with the delivery of electrical infrastructure and components—posing challenges to keeping up with the rapid pace of demand from AI developers.

Still, the trend signals a resurgence of California's role in the data infrastructure ecosystem, particularly as AI reshapes the digital landscape. PG&E’s expanding pipeline illustrates that, even amid stiff national competition, California continues to attract large-scale digital infrastructure investments, especially when it comes to powering the next generation of artificial intelligence technologies.