• 2 hours Rosneft Profits Plunge 68% as Oil Oversupply and Sanctions Pressure H1 Results
  • 2 days Annual US Crude Production Sets New Record in 2024, But Growth Pace Slows: EIA
  • 2 days U.S. Oil Output Beats Weekly Estimates in June
  • 2 days Norway Begins Peak Gas Pipeline Maintenance
  • 3 days Oil Industry Gains Ground in California Regulatory Battle
  • 3 days Oil Deal Between China and Taliban Falls Apart
  • 3 days Guyana's Oil Boom Fuels Contentious Reelection Bid
  • 3 days Strathcona Seeks to Block $5.7 Billion Cenovus Deal to Buy MEG Energy
  • 3 days Favorable Prices Prompt India to Raise U.S. Oil Purchases
  • 3 days Citgo Auction Heats Up as Bidder Sweetens Offer
  • 3 days Europe Jet Fuel Imports From Asia Hit Record High
  • 3 days Oil Prices Dip but Stay on Track to Extend Last Week’s Gains
  • 3 days EIA Delays Key Reports Following Staff Cuts
  • 3 days California Faces High Pump Prices as Phillips 66 Shuts LA Refinery
  • 3 days Exxon Sees 20% Gas Demand Growth by 2050, U.S. Nearing Records in 2025
  • 4 days Exxon Serves Up Hard Lesson in Climate
  • 4 days Repairs at Drone-Damaged Russian Baltic Fuel Port Could Take Months
  • 4 days Cargo from Sanctioned Russian Arctic LNG 2 Docks in China
  • 4 days Germany Boosts Energy Security with New LNG Terminal
  • 4 days Indian Refiners Increase Russian Crude Purchases Despite Tariffs
  • 4 days UK Regulator Investigates Drax Over Biomass Claims
  • 4 days Saudi Arabia's Oil Export Value Slumped by 16% in Q2 as Prices Sagged
  • 4 days Trans Mountain Seeks Takers for Additional Capacity
  • 4 days Petrobras Moves Closer to Amazon Oil Production
  • 4 days German Gas Traders Turn to Canadian LNG for Swap Deals
  • 4 days Oil Prices Slip on Oversupply Concerns
  • 4 days Russia’s Arctic LNG 2 Hits Record Output as Ice Conditions Ease
  • 4 days Gazprom Neft Earnings Halved In H1 On Taxes, Weak Ruble
  • 4 days Greenland’s Energy Stakes Trigger Denmark-U.S. Diplomatic Clash
  • 5 days Lower Oil Prices Dent CNOOC Earnings Despite Record Output
  • 5 days Oil Flows Set to Resume on Key Pipeline for Kazakhstan
  • 5 days UK Energy Bills to Climb as Price Cap Rises
  • 5 days Mitsubishi Abandons Three Offshore Wind Projects in Japan
  • 5 days Chinese Mining Giant Warns of Unprecedented Metal Market Risks
  • 5 days China Set to Extend LNG Import Decline
  • 5 days Nigeria Is Eager to See Petrobras Return to Its Oil Patch
  • 5 days Venezuela Deploys Ships to Oil Port
  • 5 days Goldman Sees Oil Falling Below $55 in 2026
  • 5 days China’s Industrial Profits Shrink 1.7% as State Firms Falter
  • 5 days US Oil and Products See Draw Across the Board
Leonard Hyman & William Tilles

Leonard Hyman & William Tilles

Leonard S. Hyman is an economist and financial analyst specializing in the energy sector. He headed utility equity research at a major brokerage house and…

More Info

Premium Content

Utilities, AI, and the Quiet Raid on Consumers

  • Utilities are quietly signing electricity deals with AI firms that shift major capital costs to regular customers.
  • Utilities are placing AI-related infrastructure in their regulated rate base, socializing risks and costs while offering AI firms preferential, often secretive, pricing.
  • Policymakers would be wise to force AI firms to build their own power infrastructure, shielding consumers from excessive costs.

Server room

Ok, we are cynical. The current electric utility policy environment is not exactly what you would call a level playing field, fairly balancing corporate and public interests. Quite the contrary. Right now, we have highly profitable (and politically influential) corporations facing underpowered civil servants in diminished regulatory agencies. State regulators are in a position to grant data centers and possibly other enormous users of electricity the opportunity to milk huge subsidies from unsuspecting consumers. How? By putting these vast new power-generating resources in the utility’s rate base, thereby socializing these enormous incremental costs, facilitated by pro-business politicians.

Harvard researcher (Daniel Oberhaus, “How AI Could Be Raising Your Energy Bill, ” Harvard Magazine, July/August 2025) cites more evidence that utilities are making AI power deals whose terms are not public that burden the rest of their customers. One utility plans to build several large power plants to serve a long term contract with an AI site, put the plants in their regulated rate base, but so far, no details on apportionment of costs have been released. Another large utility has been accused of offering a cut-rate deal to an AI firm with the full expectation that the rest of its customers would make up the profit differential. Stated simply, residential utility customers would be subsidizing corporate or AI  electricity usage. In addition, several AI centers announced deals in which they would take the output of existing deregulated stations. Those are perfectly above-board transactions, but not neutral to consumers who have to finance new power stations at current, relatively high prices to replace the output taken by AI. Just a guess, but we think that if AI adds 10% to system sales, it will add 30% to the fixed and capital costs of the utility, so if the AI firm gets a discount, who pays the higher costs? You guessed it.                                                    

Next, let’s consider a few wild thoughts.

For instance, might we be in an AI bubble, on par with the dot.com bubble and the power generator bubble, and way back, the bowling alley bubble? These economic or speculative bubbles burst because actual demand did not materialize to keep up with supply based on overly optimistic forecasts. As a result, entrepreneurs overbuilt or somebody came up with a better or cheaper product. “Of course”, you say, ”this time is different.” But someone always says that.

Let’s consider possibilities. First, the AI providers may have wildly overestimated electricity demand, which will leave a lot of underutilized AI data centers looking for ways to dodge those big power bills. Second, the Chinese will develop cheap, low powered AI related computer chips that make the US version look uneconomical. For example with electric vehicles, Tesla came first but Chinese auto makers have caught up fast and now offer superior products. As Andrew Carnegie supposedly and ungrammatically said, “Pioneering don’t pay.” Third, when quantum computing makes an earlier-than-expected appearance, will all those AI data centers become white elephants?

Mind you, we don’t object to building AI centers. Entrepreneurs should take chances and reap the rewards. We just object to making consumers subsidize this essentially speculative activity via their electric bills. Given the intricacies of the electric network, and cross subsidies in pricing, the divergence in cost between new and old plant, and the political pressure to give large corporations what they want via hidden extra charges in the customer's electric bill, we suggest that there is only one way to protect consumers from what AI will do to the grid. That is by requiring that AI and similar power guzzlers build their own generation and networks. Put all these new expensive generating assets behind the meter so to speak. That means build, not snatch existing stations that consumers will have to replace. That way, data centers will pay full freight and consumers will only have to deal with all the other costs increasingly burdening the utility network.

By Leonard Hyman and William Tilles for Oilprice.com

More Top Reads From Oilprice.com


Download The Free Oilprice App Today
Download Oilprice.com on Apple Download Oilprice.com on Android

Back to homepage





Leave a comment

Leave a comment




EXXON Mobil -0.35
Open57.81 Trading Vol.6.96M Previous Vol.241.7B
BUY 57.15
Sell 57.00
Oilprice - The No. 1 Source for Oil & Energy News