Trump’s data centre power ‘mandate’ goes beyond his own pledge

A separate attempt to legislate on costs encountered a Senate obstacle three weeks before Trump’s post. The House-passed Ratepayer Protection Act would require state regulators and relevant utilities to consider a cost-recovery standard for IT sites with peak electricity demand of at least 100 megawatts. Republican senator Jon Husted described the effort as a way to “turn that commitment into law”. On 30 September, a vote to end debate on whether to take up the bill received 57 votes to 43, short of the 60 required. That was a procedural defeat, not a final vote on the bill itself. [6] [7]

What a community ‘dividend’ looks like

Trump presents local payments as a route to getting projects built. His appeal comes amid opposition to AI data centres that crosses ideological lines, according to Reuters. He said previously unpopular centres had become “extremely popular” after financial deals, but named no project as an example. [1] [12]

A recent proposal in De Soto, Kansas, illustrates the sort of benefit he advocates, although he did not mention it. On 9 October, Digital Realty announced a $20 million commitment towards household utility and property-tax credits alongside its Astra North development. The company said it was working with the city to finalise credits of up to $1,000 per household annually for ten years, based on current population estimates. Further details were to be confirmed. [8]

The company’s announcement quoted De Soto mayor Rick Walker welcoming the investment in residents. That establishes a supportive response from a local official to a proposed benefit. It does not demonstrate the shift in residents’ opinions Trump described, and the projects behind his assertion remain unnamed. [1] [8]

Lower prices depend on supply and contracts

Financial gains for electricity customers are possible. In its July announcement, the White House cited NiSource arrangements with Amazon and Alphabet that it said were expected to return at least $1.4 billion to Indiana customers over 15 years. That timetable makes the figure a long-term projection, rather than an immediate reduction in household bills. [4]

Historical research also offers support for the possibility of lower rates. A preprint by Asa Watten, John Bistline and Geoffrey Blanford, affiliated with the Electric Power Research Institute and Watershed, estimates that data centres kept average residential electricity rates modestly lower than they would otherwise have been in 2015–2024. The authors explain that sustained demand can spread fixed power-system costs across more electricity sales. They warn, however, that future supply constraints could reverse the effect. [9]

The more recent price picture is less reassuring. A June literature review from Columbia University’s Center on Global Energy Policy, produced with support from Google.org, reports that average US residential electricity prices rose at more than twice the rate of inflation in 2025. It identifies several sources of pressure, including network investment, disaster recovery and fuel-price volatility, rather than demand alone. It also warns that data centres can be built faster than the generation and transmission infrastructure needed to serve them. [10]

The practical test is whether additional supply arrives in time and whether developers, rather than households, carry the costs of serving it. The published pledge allows bought-in power to meet that aim; Trump’s claim of compulsory self-generation goes beyond its terms. [2] [10]