Across the U.S., there’s a significant push to build new data centers, which require vast amounts of electricity. As demand surges, utilities are scrambling to expand their generation and infrastructure to accommodate these energy-hungry facilities while finding ways to protect other customers.
Experts suggest that several challenges could impact how utilities manage this risk, potentially leading to stranded investments. The power industry typically plans for decades in advance. However, the spike in demand for data centers, particularly influenced by booming AI technologies like ChatGPT, has arrived rapidly.
These AI-focused data centers have distinctly different energy profiles compared to traditional ones, making them challenging to replace once integrated into utility frameworks. Predicting their future energy needs is complicated due to the fast-paced changes in both AI and its underlying hardware.
Some analysts liken the uncertainty surrounding AI-related data centers to the dot-com bubble. The aftermath of that burst left behind enormous amounts of unused fiber-optic cables, which were eventually repurposed. In contrast, the current scenario doesn’t have a clear backup plan for underutilized energy infrastructure.
There are concerns about the sustainability of businesses that rely heavily on these AI services. Much of the financial backing for newer cloud companies hinges on major players like Microsoft and NVIDIA. If these companies falter, it could have disastrous ramifications across the whole sector.
The potential financial risks extend to utilities as well. With large customers promising significant energy payments, their financial stability is crucial. If a large company were to go bankrupt, the financial burden could shift back to the remaining customers.
Utilities are adopting new strategies, including long-term contracts and large load tariffs, to mitigate risk. For instance, American Electric Power recently initiated a study fee for large load proposals, drastically reducing its forecasts for future demand.
Innovative pricing structures may help stabilize the system, but concerns remain about whether smaller utilities can navigate these changes effectively. Some, like the Northern Virginia Electric Cooperative, expect to see data center customers making up an overwhelming majority of their sales in the coming years.
Overall, as demand for data centers continues to grow, utilities will need to adapt swiftly to address these evolving challenges while ensuring that the system remains balanced and efficient.

