Christopher Hailstone joins us to navigate the complex intersection of the artificial intelligence revolution and American energy stability. As an expert who has spent decades analyzing grid reliability and the shift toward modern electricity delivery, he provides a grounded perspective on why the current rush to power massive datacenter campuses is more than just a trend—it is a fundamental restructuring of our national infrastructure. In this conversation, we explore the looming capacity gaps, the surprising resurgence of traditional fuels to maintain system resilience, and why the fear of overbuilding might be the biggest mistake policy makers could make in the current economic climate.
With datacenters projected to drive 75 percent of all US power demand growth over the coming decade, how do we reconcile the risk of overbuilding if the current AI implementation struggles to produce positive outcomes?
It is a delicate balance for any utility provider, but the reality is that the risk of underinvestment is far more catastrophic than building a bit too much capacity. Currently, we are seeing about 71 percent of organizations reporting negative outcomes from their initial AI implementations, which naturally raises eyebrows about a potential bubble bursting. However, even if the AI frenzy cools down, the infrastructure we are building today—the transmission lines, the substations, and the generation plants—is not going to sit idle or become a stranded asset. This hardware will immediately be swallowed up by the broader electrification of our economy, from the rapid rise of electric vehicles to the industrial shift away from legacy processes. We are looking at a need for 30 gigawatts of additional power every single year just to keep pace, including roughly 20 gigawatts for IT equipment and the remainder for cooling and distribution, making this foundation essential for grid resilience regardless of whether a specific software sector booms or busts.
The projected nationwide capacity gap is reaching up to 55 gigawatts by 2030 as older plants retire. What measures are being taken to keep the lights on while we wait for new infrastructure to come online?
We are currently in a high-stakes “bridge” period where we have to keep the old systems running much longer than originally planned to avoid a total supply failure. While we have about 40 gigawatts of spare dispatchable capacity and 100 gigawatts of committed new capacity in the pipeline, we are also losing 50 to 75 gigawatts of coal and gas-fired steam capacity to scheduled retirements. To prevent a total shortfall and maintain grid reliability, there has been a significant push to bring mothballed plants back to life and extend the operational years of existing facilities. For instance, the administration recently authorized up to $500 million in funding to keep 13 coal-fired plants operational, specifically to bolster the resilience of our critical energy infrastructure. These are not permanent solutions, but they provide the essential temporary relief needed while we wait for the 120 gigawatts of new demand growth to be met by more sustainable, modern sources.
Why are we seeing such a massive shift toward on-site power generation for these campuses, and do you expect this to be a permanent fixture of the industrial landscape?
The shift is driven by a simple, frustrating reality: the public grid cannot provide connections to these massive facilities fast enough to meet their operational timelines. Developers are tired of waiting years for a grid connection, so they are taking matters into their own hands with gas turbines, fuel cells, and large-scale battery storage. Our surveys show that nearly 60 percent of power sector leaders expect these datacenters to keep their on-site generation permanently, even after a standard grid connection becomes available. It provides a level of “five-nines” reliability that a strained public grid simply cannot guarantee right now. About 64 percent of those planning these on-site setups are relying on natural gas because it is cost-competitive and can be deployed much faster than waiting for a regional utility to upgrade its local transmission infrastructure.
Looking toward the next decade, where do you see the most significant contributions coming from in terms of new generation technology, specifically regarding nuclear and geothermal energy?
While there is a lot of capital and buyer interest flowing into small modular reactors (SMRs) and next-generation geothermal energy, they aren’t the immediate heroes of this capacity story. We don’t expect those emerging technologies to have much of an impact on our system capacity until at least the middle of the 2030s. Between now and then, the heavy lifting in the nuclear sector will come from extensions and upgrades to our existing fleet of plants rather than brand-new builds. Beyond 2030, we will see a much stronger marriage between solar and battery storage as they become more cost-effective. Solar is incredibly fast to deploy, but it has to be part of a diversified energy system that includes natural gas to ensure we aren’t left in the dark during peak demand periods when the weather doesn’t cooperate.
What is your forecast for US energy infrastructure over the next five years?
I anticipate a period of intense, almost frantic construction that will fundamentally redefine the American industrial landscape. We are currently watching energy consumption in the datacenter sector grow by 26 percent this year alone, and that momentum is going to force a massive modernization of our transmission networks. We will see a “dual-track” energy system emerge where large industrial users operate their own microgrids while the public grid focuses on incorporating the massive amounts of solar and storage needed to fill the 30 to 55 gigawatt gap. It will be a bumpy ride with potential localized shortages as we phase out older plants, but the end result will be a significantly more robust and resilient power system that can handle the massive loads required for the high-tech future.
