Data centers could consume somewhere between 9.5% and 15.3% of all US electricity by 2030, up from just 4.7% in 2024. That is a staggering jump, and it has ratepayers and regulators bracing for a wave of higher electric bills. But a new report from Columbia University pushes back on the idea that soaring prices are inevitable. The researchers argue that smarter use of the grid we already have can blunt much of the cost pressure, provided utilities are willing to change how they operate.
Their prescription centers on Grid-Enhancing Technologies, or GETs. These are not science fiction; they are practical, relatively cheap tools like dynamic line ratings, advanced conductors that carry more current, and power-flow controllers that route electricity around congestion. Paired with expanded demand response paying large users to dial back consumption during peak hours, these technologies can effectively create new capacity without stringing a single new high-voltage line. In a world where building transmission can take a decade, squeezing more out of existing assets is enormously valuable.
The conventional utility response to load growth is to build new lines, new substations, and new generation, all of it recovered from ratepayers over decades. That approach is slow, expensive, and increasingly unpopular. The Columbia analysis suggests there is a cheaper, faster path that could keep near-term prices in check while the bigger infrastructure questions get sorted out. For households worried that the AI boom will quietly inflate their monthly bill, that is genuinely good news, but only if the tools actually get deployed at scale.
The researchers point to a structural problem in how utilities are regulated. Under traditional cost-of-service models, utilities earn a guaranteed return on capital they invest. This means they are financially rewarded for building expensive new hardware, but not for the cleverness of optimizing what already exists. That incentive structure quietly tilts the entire industry toward steel and concrete over software and sensors. Fixing it requires regulators to rethink how utilities make money, which is a heavier lift than installing the technology itself.
Keep an eye on state public utility commissions, where the real decisions about cost recovery and performance incentives get made. Several states are already piloting "shared savings" models that reward utilities for efficiency rather than spending. If those spread, GETs adoption could accelerate quickly. Watch whether the big data center developers themselves embrace flexible operations, since their willingness to curtail during peaks is a huge variable. The technology to keep the AI power bill manageable largely exists. The open question is whether the incentives will catch up in time.
SHARE THE BREW
If Energy Brew made you smarter before your second coffee, send it to someone who reads power markets for fun or pretends not to.
Share the brief and help build the daily energy intelligence habit.
