While AI is developing rapidly, its side effects as a "power-guzzling monster" are beginning to emerge, even becoming a focal point of political maneuvering. On January 12, local time, US President Trump posted on his social media platform Truth Social, launching a scathing attack on tech companies for building massive data centers, which he said had caused soaring electricity bills for American households, and demanding that these companies must be "self-sufficient."
Microsoft responded quickly, promising to cover all costs of the relevant power infrastructure construction and ensuring that these costs are not passed on to ordinary community residents.
Trump: I absolutely do not want Americans to pay higher electricity bills for AI.
In his post, Trump pointed out that the average monthly utility bill for American households had increased by more than 30% at one point, and blamed the previous Biden administration and the recent AI boom. With the large-scale construction of data centers, the power supply is beginning to face enormous pressure.
While Trump emphasized that the United States must maintain its leading position in artificial intelligence, and that data centers are a key infrastructure component, he also stated that he "absolutely does not want Americans to pay higher electricity bills because of data centers." He demanded that large technology companies be accountable to the American people and announced that the government is working with technology companies to announce several initiatives in the coming weeks, with Microsoft being the first company to make a commitment.
Learning from the Wisconsin lesson? Microsoft swiftly offers five major promises.
Following Trump's tweet, Microsoft President and Vice Chairman Brad Smith delivered a speech titled "Building Community-First AI Infrastructure."long essayThey announced the launch of a new plan and made five commitments.
The most crucial point is that Microsoft will "pay out of its own pocket." Microsoft has pledged to require utilities to set electricity prices at a level sufficient to cover data center costs, meaning that Microsoft will absorb all the additional electricity costs instead of passing them on to local residents' bills.
In addition, Microsoft's specific plans also include:
• Transparent cooperation:Early and transparent cooperation with power companies ensures that electricity is supplied to the grid when needed.
• Improve efficiency:Improve the energy efficiency of data centers through technological innovation.
• Policy Initiatives:Promote necessary state and national policies to support communities’ access to affordable and stable electricity.
Microsoft's proactive stance stems largely from its painful experience in 2025. Last year, Microsoft planned to invest $7 billion in Wisconsin to build a 0.9GW data center, but the project was forced to be canceled in October due to strong opposition from local residents over electricity cost sharing.
The key to the AI race: electricity
The backdrop to this political and commercial struggle is the growing "AI power shortage anxiety" in the United States. The International Energy Agency (IEA) predicts that electricity demand from U.S. data centers will more than triple over the next decade, putting enormous pressure on the aging and chronically under-stretched U.S. power grid.
Microsoft CEO Satya Nadella previously admitted that the biggest bottleneck facing the AI industry is not an overabundance of computing power, but rather the inadequacy of power infrastructure. "If there is no support, chips will just sit in warehouses and be unusable."
For the Trump administration, this is a hot potato that it has to deal with. Although he has been pushing the AI industry and easing restrictions, rising inflation and utility costs have already sparked discontent among many voters. If the electricity bill issue is not resolved, it could impact the results of the midterm elections in November.
Analysis of viewpoints
In the past, tech giants were accustomed to enjoying relatively cheap industrial electricity or negotiating tax and utility concessions in exchange for promises of "job creation." However, as AI data centers grow to gigawatt (GW) scale, their electricity consumption has become large enough to impact the stability and price of residential power grids. Microsoft's response has been swift and pragmatic; after their setback in Wisconsin, they have a deep understanding of the "not in my backyard" (NIMBY) effect.NoteThe power of ).
Note:The NIMBY (Not In My Back Yard) effect is a social phenomenon in which residents oppose the construction of public facilities (such as landfills, nuclear power plants, funeral homes, and base stations) near their homes by the government or businesses that may have negative impacts (including pollution, noise, falling property prices, and health risks), even if these facilities are beneficial to society as a whole.
This establishes a new rule of the game: the future AI computing power competition is essentially an energy competition. Want to build a data center? Prepare your own power plant or pay exorbitant electricity bills. This will be a huge cost burden for companies like Google, Amazon, and Meta, but it may also accelerate the adoption of supermassive nuclear power (SMR) and green energy technologies in the tech industry.
