What’s Driving Higher Power Bills in the U.S.? (Full Transcript)

Grid upgrades, material shortages, tariffs, extreme weather and AI data-center demand are pushing electricity bills higher—often faster than usage.
Download Transcript (DOCX)
Speakers
add Add new speaker

[00:00:00] Speaker 1: This is my electric bill covering October of 2025, where I used 633 kilowatts. And here's my bill from just a year before, where I used 602 kilowatts of power. And despite only using 5% more electricity, my bill went up more than 12%. You might relate. Some states have seen their prices go up 5%, 10%, even 15%. And it'd be easy to just blame data centers for this. But prices are increasing places that have just a few of them, and not always where they have a lot of them. So why are electricity bills going up? Your bill, like my bill, is calculated from these two main categories. Supply, so the actual electricity, and the delivery of that electricity. Which is simply the cost of these, all the transformers and towers and wires that make up the grid. The U.S. grid is old. Electric companies reported that most of their investments were replacing or adapting the current infrastructure. And if you're in a place like California, the cost of replacing what's been damaged from wildfires, and what needs to be quickly replaced to prevent wildfires, is high. One study found it's equivalent to a 4 cents a kilowatt increase to people's bills. Which doesn't seem expensive, but remember my small house last October used 633. It's why California has the highest cost of electricity in the continental U.S. Hurricane damage in Florida and North Carolina have increased electricity costs there too. But really, these infrastructure costs are up everywhere, because the costs of the materials have really increased. Things like switchgears, all that power cable and wire, and transformers. Not only are they more expensive, they're harder to actually get. In 2019, it took 3 to 6 months from ordering to receiving one. By 2023, it was 1 to 3 years. A lot of this is because of COVID supply shocks, especially with the raw metal materials. And it hasn't been helped by the war in Ukraine.

[00:01:59] Speaker 2: The Russians targeted basic infrastructure like electrical poles, transmission wires. There was a concerted effort led by the federal government to try to ship transformers to Ukraine to try to replace what the Russians were destroying. Of course, it led to a shortage. And whenever you have a shortage, you know, the price goes up.

[00:02:18] Speaker 1: And Trump's 25% tariffs on many metals went into effect here. They've since been raised to 50%. We don't even see that full effect on this chart yet.

[00:02:28] Speaker 2: When tariffs are placed on steel, copper, aluminum, they drive up the cost of utility assets that have to use steel, copper, aluminum. Three of the most common raw materials in the utility industry.

[00:02:42] Speaker 1: So the increased cost of those grid repairs and expansions are paid for by us. And so is the cost of the increased demand for power. This is where we talk about data centers. Researchers looked at the increase of electricity demand and compared to the increase in price, and found that the more demand, the lower the price of electricity, not higher. What happens is these costs for the grid are divided up among all the customers, like you and me. So adding a new development or a new shopping center would require a minimal amount of new infrastructure and would mean more people to split that original cost up with. And as long as it doesn't use so much more electricity that they need to build new power plant, it wouldn't make energy prices go up either. Hence, adding more people to the grid, the lower the price. But this study was between 2019 and 2024, before the data center boom in 2025. The problem with data centers is they aren't just a new customer, they're like a new town. And while their grid needs are minimal, they take up nearly all that extra power, making the power more expensive. The average house uses less than 11,000 kilowatts a year. A single AI data center can consume more than 800 gigawatts. In all, data centers used 183 terawatts in 2024 in the U.S. That's more than the entire state of New York used. Experts told me that my bill in Maryland went up partially because of those increased costs to repair the grid and partially because of data centers. And it could be the same for you too, even if you don't live near one. The U.S. has many different companies that operate the grid. Texas famously has its own. I'm on one that also includes Virginia up to Pennsylvania and even over to Chicago. We're all in the same system. So when data centers increase demand in Virginia, folks in Ohio and New Jersey and I share the cost. Now, every state also has its own regulations over these utilities, and they each have different systems for calculating that cost of energy generation.

[00:04:39] Speaker 2: What you're paying for power very much depends upon your individual state and the policies that your state has adopted.

[00:04:47] Speaker 1: And it's why there's many different proposed solutions to the data center problem. Because data centers are only growing. It's estimated that they could use between 325 to 580 terawatts by 2028. Some states are freezing the amount bills can increase. Others are proposing that data centers be charged a different electricity rate or generate their own power completely. Some experts say that with enough new power generation, it can help keep costs for everyone down. But the costs to build new power keep increasing, as do the costs to expand the grid.

[00:05:20] Speaker 2: It's going to get a lot worse. AI-driven data centers are taking huge amounts of power. We've got to find a way to provide generation for the data centers. At the same time, be fair and just to the millions and tens of millions of residential consumers. And that's the challenge we face right now, making sure those costs are fairly allocated.

ai AI Insights
Arow Summary
Electricity bills are rising faster than usage due to higher delivery (grid) costs and, in some regions, surging demand from AI data centers. The U.S. grid is aging and requires expensive repairs and upgrades, with costs worsened by wildfire and hurricane damage. Material and equipment prices have jumped and lead times have stretched because of COVID-era supply shocks, Ukraine-related transformer shortages, and tariffs on metals like steel, copper, and aluminum—costs ultimately passed to ratepayers. While added customers can sometimes lower per-customer costs by spreading fixed grid costs, large data centers can behave like “new towns,” consuming most available incremental power and pushing up generation needs and costs across shared regional grids. State policies and utility regulations strongly affect what consumers pay, and proposed solutions include rate caps, special data-center rates, or requiring data centers to self-generate power, but both generation and grid expansion are getting more expensive.
Arow Title
Why Electricity Bills Are Rising: Grid Costs, Supply Shocks, and Data Centers
Arow Keywords
electricity bills Remove
utility rates Remove
grid infrastructure Remove
delivery charges Remove
transformers Remove
supply chain Remove
tariffs Remove
steel Remove
copper Remove
aluminum Remove
wildfires Remove
hurricanes Remove
Ukraine war Remove
data centers Remove
AI Remove
electricity demand Remove
regional grids Remove
state regulation Remove
rate design Remove
cost allocation Remove
Arow Key Takeaways
  • Bills can rise significantly even with small usage increases because delivery/grid costs are growing.
  • Aging infrastructure and climate-related damage (wildfires, hurricanes) drive major upgrade and replacement spending.
  • Equipment and material costs are up and delivery times for key components like transformers have lengthened.
  • COVID supply shocks, Ukraine-related transformer diversion, and metal tariffs contribute to higher utility capital costs.
  • Those higher capital costs are typically recovered from customers through rates.
  • Demand growth doesn’t always raise prices, but very large new loads (AI data centers) can force new generation and strain available power.
  • Because grids span regions, data-center-driven costs can be shared across multiple states.
  • State regulations and rate-setting methods strongly influence consumer bills.
  • Potential policy responses include rate caps, separate data-center tariffs, or requiring data centers to provide their own generation.
  • Balancing rapid load growth with fair cost allocation to residential customers is the central challenge.
Arow Sentiments
Neutral: Explanatory and analytical tone focused on cost drivers (infrastructure, materials, policy, and demand) with caution about worsening impacts; minimal emotive language beyond concern about future increases.
Arow Enter your query
{{ secondsToHumanTime(time) }}
Back
Forward
{{ Math.round(speed * 100) / 100 }}x
{{ secondsToHumanTime(duration) }}
close
New speaker
Add speaker
close
Edit speaker
Save changes
close
Share Transcript