Professional sport teams have adapted to the market supply/demand balancing act by anchoring their brand to structured elements of their business model.  Take soccer for example, Major League Soccer (MLS) now has a layer in its structure called ‘Next’.  This layer did not exist a few years back but now has the attention of every player and club in respective metro areas with a team. If you live in a more established structure (have a team of many years) and you travel to a city that is in the early phases, you can see the vision and issues that will occur along the way. In the energy space, specifically electricity, the same type of layering is taking place with load growth being the latest in the supply/demand balancing conversations across all regions in North America.  In this piece, we will focus on California’s Next (load growth) as it will be a topic of discussion in the months/years to come.

Projecting electricity demand growth in CAISO at this time comes with a high degree of uncertainty due to three primary factors: first is the underlying demand itself–data centers, a major source of projected load growth as they themselves don’t know how much power will be needed; second, is the ability to interconnect large new loads to the electricity grid; and third is the ability to add sufficient generation supply to meet this new demand. This uncertainty is highlighted by the March 2025 California Energy Commission (CEC)’s California Energy Demand Forecast.  This recent forecast is 3.2 GW above the prior forecast for 2028, and 6.1 GW above the prior forecast for 2031.

The largest increase in load can be attributed to data center load growth[1] and a reduction in behind-the-meter solar by half. The figure below compares the new statewide CEC load forecast to the prior one for 2031. The graph depicts the CEC 2024 load forecast (“Base Load”) and the 2025 load forecast (“High Load”) as implemented in the Energy GPS production cost model.

Figure 1 | Average Hourly California Demand, Base vs. High Load 2031

Energy GPS uses an Aurora-based production cost model (PCM) to model the future conditions throughout the western interconnect. Our current model relies on the prior CEC load forecast. It is difficult to know the timing of load additions compared to supply additions. It is also difficult to know how much of this increase in demand has already been planned for by its load serving entity. There is a high probability that this increase in demand will outpace the supply additions.

[1] The CEC releases an hourly forecast file that includes CAISO load breakdowns by category. A statewide version of the equivalent breakdown does not seem to be available. By 2031, there is an additional 2.6 GWs of data center load forecasted (up from zero in the 2024 forecast), as shown in Figure 2. This is the first year for data center load growth to be explicitly included in the CEC forecast and was the largest category for load increase.