Most of what we pay for isn’t the power itself — but the aging network that delivers it
By Doug Kelly
WASHINGTON, D.C. (Texas Insider Report) — As electricity bills rise, Americans have heard a familiar refrain: artificial intelligence (AI) and data centers are to blame. From media headlines to the campaign trail, AI has become a convenient scapegoat. But a major new study from the Lawrence Berkeley National Laboratory and The Brattle Group tells a very different story.
While AI and data centers have increased electricity demand, the real culprit behind higher costs is decades of under-investment in America’s power grid — the transmission lines, poles, transformers, and other essential infrastructure that keep electricity flowing.
Decades of Energy Grid Under-Investment & Neglect
As a recent Washington Post investigation found:
“U.S. utilities haven’t been on top of replacing power poles and lines in the past, and are now trying to catch up.”
That catch-up is proving expensive, and the costs are being passed directly onto consumers.
In fact, over the past two decades, while the cost of power generation has fallen by about 35%, the costs of transmitting that electricity across an aging grid have nearly tripled, and distribution costs have more than doubled.
These grid shortfalls—not AI—are what’s really driving up electricity bills.
Our energy infrastructure picture is stark: In 2025 the American Society of Civil Engineers gave U.S. energy infrastructure a D+ grade, down from a C- in 2021. More than 70 percent of transmission lines are over 25 years old, while many grid components built in the 1960-70s are now nearing the end of their 50- to 80-year design lifecycles.
Even more alarming, construction of new high-voltage transmission lines is far behind what’s needed. Federal data show the U.S. built just 888 miles of new 345 kV+ high-voltage transmission in 2024 — compared to the roughly 5,000 miles per year experts say are required to keep up with demand and maintain affordability.
The table below spotlights the past decade of decline in transmission line investment.
This failure to invest matters because high-voltage lines dramatically reduce the cost of moving power. One analysis found that the cost per MW-mile of a 765 kV line is less than one-third that of a 345 kV line.
So not only are consumers operating on an inefficient system, but they also now face higher bills as utilities scramble to replace decades of deferred investment.
While America Scapegoats, China Supports Its AI Innovators
While some policymakers blame AI and data centers for higher power bills, China is doing the opposite — it’s building more energy and backing its AI companies.
Beijing invested roughly $300 billion more in energy infrastructure than the U.S. in 2025 alone — and now generates twice as much electricity as America. That energy edge is central to China’s AI ambitions and its push to dominate the globe’s digital domain. In short: China is not just constructing the backbone of the AI era — it is subsidizing it.
For example, Reuters reported that Chinese authorities are offering cheap electricity and power subsidies to major tech firms, including ByteDance, Alibaba, and Tencent, to offset the higher costs of running data centers on domestic AI chips. Local provincial governments have expanded these incentives to keep China’s AI infrastructure push on track.
Meanwhile, in the United States, state legislators have introduced more than 1,100 bills related to AI, potentially creating a blizzard of red tape.
This contrast — “the U.S. blames while China builds” — should alarm every policymaker in at every level. While utilities in the U.S. scramble to replace 75-year-old lines, China is constructing the AI infrastructure of the next century.
What the Study’s Data Actually Found
That’s what makes the new Lawrence Berkeley–Brattle findings so striking. While AI and data centers have contributed to higher demand, that demand hasn’t been what’s driving electricity prices up as of yet.
In fact, the researchers found that states with more electricity demand — the very places hosting AI data centers — often experienced lower prices overall.
- North Dakota, for example, saw electricity demand jump nearly 40 percent, driven in part by a surge of new data centers, yet its inflation-adjusted prices fell by about three cents per kilowatt-hour (kWh).
- Virginia, the nation’s leading data center hub, recorded a 14 percent rise in demand and a one-cent drop in prices.
- California, by contrast, saw a small decline in demand but its prices increased by more than six cents per kWh.
How can increased demand lead to lower prices? Because electricity doesn’t follow economics 101. Most of what we pay for isn’t the power itself, but the aging network that delivers it. When more customers share those fixed costs, rates can fall, not rise, meaning more electricity demand can actually lower prices.
The Way Forward
We need to get to work right now. If Congress wants to keep America competitive and keep energy prices affordable, its first job is to codify and fund President Trump’s AI Action Plan. The plan jump-starts new transmission projects, expands domestic energy supply, and cuts regulatory red tape, ensuring we have the power to lead in the AI age.
Winning in AI isn’t just about chips and code. It’s also about whether we can power the future. And right now, that means catching up on the one thing we’ve neglected for far too long: the grid itself.
Doug Kelly is CEO of the American Edge Project. For over 20 years, he has successfully led three separate large-scale transformation efforts in technology, organization building, and fundraising, and has also held leadership roles in Presidential, Gubernatorial, and other statewide campaigns.
