Across the country, utilities are scrambling to meet unprecedented power demand from artificial intelligence, cloud computing, and digital manufacturing. Behind the sleek promise of the AI revolution lies a simple truth: all that intelligence runs on electricity. And according to Bank of America’s latest analysis of utility spending, the bill for powering the digital economy is growing fast — and beginning to show up on household balance sheets.

The report projects U.S. utility capital expenditures will reach $174 billion by the end of 2024 — the highest in history.

Roughly 42% of that spending is going toward transmission and distribution upgrades, the backbone of the grid. The surge is driven largely by rising electricity demand from data centers, which now consume 6%–8% of total U.S. generation and could rise to 15% by 2030.

The issue isn’t just how much utilities are spending — it’s who’s paying for it. Wholesale power prices are expected to increase 19% between 2025 and 2028, while the average national electricity rate has already climbed 2.7% year over year. As utilities pour money into new infrastructure, much of the cost is likely to pass through to ratepayers. The cloud, it turns out, is anything but free.

AI’s Energy Appetite

Bank of America’s report highlights how artificial intelligence, machine learning, and the broader digital economy are rewriting electricity forecasts. Roughly three-quarters of the nation’s top 35 utilities have reported a sharp rise in demand from data centers. These facilities — windowless warehouses of servers and cooling systems — already consume more power than some mid-sized countries.

For utilities, this marks the end of a 20-year era of flat demand. After decades of efficiency gains and slow load growth, electricity use is now surging — fueled by AI, electrified transport, and industrial reshoring.

Meeting this demand requires rapid investment in substations, transmission lines, and generation capacity. Bank of America calls this the start of “a fundamentally new electricity landscape.”

But the geographic footprint of this growth is uneven. States like Georgia, Texas, and Virginia have become magnets for hyperscale facilities, while regions like the Midwest are seeing clusters of new projects tied to manufacturing and chip production. That uneven demand is forcing utilities to reimagine resource planning in real time.

A Chain Reaction of Costs

Bank of America’s analysis shows a historic wave of utility spending, with capital outlays projected to hit $174 billion this year — up from pandemic-era averages near $120 billion. More than two-fifths of that is going to transmission and distribution, traditionally the most expensive and time-consuming parts of the system to build.

Supply chain bottlenecks are compounding the problem. The lead time for critical equipment like transformers has more than doubled since 2021 — from about 50 weeks to 120–210 weeks for larger units. Meanwhile, extreme weather costs reached $53 billion between January and August 2024, straining both infrastructure and finances.

The result: utilities are paying more to maintain reliability, often borrowing to finance projects while waiting for regulatory approval to recover costs.

That lag time means consumers start paying before they see any benefit. It’s a grid investment paradox — vital for modernization, but immediately inflationary.

Big Tech’s Boom, Everyone’s Bill

Utilities are experimenting with new rate structures to prevent household customers from absorbing the costs of Big Tech’s expansion. Some are proposing “large load” tariffs or “clean transition” rates that charge data centers more directly for the infrastructure they require. Others, still bound by traditional rate design, spread costs across the entire customer base. In markets where regulators move slowly, that approach effectively socializes the cost of powering private industry.

Bank of America notes that rate case approvals often trail load growth, creating a gap between what utilities spend and what they can recover. The difference is made up later — usually through higher residential rates.

It’s an energy version of trickle-down economics: Big Tech’s growth drives grid investment, but the near-term costs ripple through to everyone else.

The Bottom Line

Artificial intelligence may be virtual, but its energy appetite is very real. Bank of America’s report paints a clear picture of an industry entering a high-spending era, where utilities must modernize faster than regulators can respond.

For consumers, the hidden price of powering the cloud is already showing up — as a slow, steady climb in the average household’s monthly bill.