MW Trump's crusade against Big Tech's energy spending highlights a problem with no easy solutions
By Britney Nguyen and Christine Ji
'It's not something that can be solved overnight': AI has overwhelmed the electricity grid, and everyone from average Americans to Trump is outraged
In the race to build artificial intelligence, Big Tech companies have spent record-breaking amounts of money on building data centers. Now, they might end up spending even more as they face unprecedented federal intervention along the way.
To compete in the AI revolution, Big Tech companies are focused on building out as much AI data-center infrastructure as possible. But now hyperscalers like Microsoft $(MSFT)$, Meta Platforms (META), Google $(GOOGL)$ $(GOOG)$ and Amazon.com (AMZN), are facing major obstacles, including ordinary Americans fed up with the rising cost of electricity and a president increasingly focused on affordability in the months leading up to the midterms.
With energy appetites rivaling those of small cities, data centers have overwhelmed America's power grid and contributed to spiking electricity prices. On Friday, the Trump administration called for the nation's largest grid operator, PJM Interconnection, to hold an emergency auction for wholesale electricity for Big Tech companies to bid on 15-year contracts for new power plants. Contracts from the auction would support $15 billion in construction of new power capacity, Bloomberg reported, and require companies with contracts to pay for power whether they use it or not.
But experts on power generation and transmission told MarketWatch there are no easy solutions that address Big Tech's thirst for power and the tensions it creates with consumers' concern about the cost of living.
"Under President Trump's leadership, the administration is leading [an] unprecedented bipartisan effort urging PJM to fix the energy subtraction failures of the past, prevent price increases, and reduce the risk of blackouts," White House spokesperson Taylor Rogers told MarketWatch. "Ensuring the American people have reliable and affordable electricity is one of President Trump's top priorities, and this would deliver much-needed, long-term relief to the Mid-Atlantic region."
Jeffrey Shields, a PJM spokesperson, said the operator is currently reviewing the principles laid out by the White House and a group of state governors from the PJM area at the meeting on Friday, in emailed comments. And following "a multi-month stakeholder process," Shields said the PJM's board has made a decision on integrating large loads to the grid.
"The Board has been deliberating on this issue since the end of that stakeholder process," Shields said. "We will work with our stakeholders to assess how the White House directive aligns with the Board's decision."
The Trump administration's recent initiative "is not likely to have a material impact on near-term electricity prices," Josephine Tucker, JLL's Americas head of energy advisory and sustainability, told MarketWatch over email. "It could take a few years to see a rebalancing of prices for the PJM market"
Microsoft, for example, pledged to pay a premium for electricity to cover the cost of new infrastructure and forgo local tax breaks, but the model may not be replicable for smaller operators with fewer financial resources, and doesn't address supply chain issues and delays. Solutions like upgrading the power grid, generating electricity onsite at data centers and nuclear power will take years to implement, experts say.
Trump's new policy proposal also comes after most of the riches of the AI revolution and its associated stock boom have risen to the top. "The gains from AI [have] accrued mainly to the wealthy," Jesse Kuri, a geopolitical strategist at BCA Research, told MarketWatch. Kuri and some others see Trump's position on utilities, as well as his recent stances on credit-card rates and corporations buying single-family homes, as a strategy to gain support for the upcoming midterm elections.
No easy solutions
Setting up a new data-center often requires incremental generation capacity, new substations or upgrading transmission lines, Rick Pederson, chief strategy officer at Bow River Capital, told MarketWatch in emailed comments.
The Trump administration's move to make AI giants pay up is the latest example of increased state intervention in markets, especially in the realm of AI, Juan Correa, chief strategist of global asset allocation at BCA Research, said. He pointed to the U.S. government's strategic investment in Intel last year as another example.
"The United States has enormous needs for infrastructure capex for various economic, technological and geopolitical reasons," Correa told MarketWatch by email. "The problem is that the public sector is starting to run out of fiscal space to finance these projects [...] The administration is thus pushing the private sector to undertake these projects through various avenues."
The U.S. electricity grid is indeed in dire need of an upgrade. After two decades of nearly flat demand, it's ill-equipped to handle the massive needs of data centers, which now consume 5% of the nation's total power.
But it could take up to a decade before more power is widely available. Shortages of component parts and metals have dragged out construction and permitting timelines, Tucker said. On top of that, PJM already has an existing backlog of projects awaiting interconnection, she added.
"It's not something that can be solved overnight," Matt Landek, global division president of data centers and critical environments at JLL, told MarketWatch. "You can't just turn around and upgrade the grid in a 36 month cycle."
That means no matter how much Big Tech pays in the meantime, the U.S. will still face electricity scarcity. "The rapid growth and the rapid request for grid upgrades is going to continue to be a struggle for the next five to 10 years," Landek added.
To complicate matters further, the exact impacts of data centers on electricity costs are still unclear. An October report from The Economist made the case that there's no association between new data-center builds and recent spikes in power bills. In fact, massive industrial users could help lower rates by spreading the grid's fixed maintenance costs across a larger base of customers.
How can data centers pay their 'fair share'?
For Big Tech, effectively addressing public criticism and gaining more support from local communities will be critical to moving forward in the AI race. As a result, Matt Sallee, senior portfolio manager at Tortoise Capital, believes the hyperscalers will be eager to comply with Trump's new demands.
"This whole thing is really about speed-to-market, and it's basically a strategic imperative that they roll out AI even if they're not sure how they're going to earn returns on it just yet," Sallee told MarketWatch. The hyperscalers are looking to establish their footprint as quickly as possible, he said, and if the gating factor is access to electricity, they will likely be willing to pay a premium to the market price.
Going forward, Landek anticipates that the cost of power will be factored into capital raises for AI projects.
Although Sallee thinks tech companies will be cooperative with Trump, he pointed out that the decision really comes down to the states and regulatory commissions who are responsible for ensuring reliable electricity supply at a reasonable price. Local governments also have to approve where data-centers are built, and if large loads are connected to the grid.
One way to ensure data centers pay their "fair share" of electricity costs is through large-load tariffs, Sallee said. He described it as a customized electricity rate structure for large customers, typically those who will need more than 50 megawatts of power. The approach could ensure existing grid customers don't end up subsidizing large electricity loads and paying for infrastructure updates needed by the larger customers.
Another solution is behind-the-meter power, which involves data-center contracts for on-site electricity generation in lieu of connecting to the grid. Behind-the-meter options such as gas turbines or wind and solar power could be an interim solution for companies waiting to connect to the electricity grid, according to Sallee.
With behind-the-meter power becoming "increasingly relevant for hyperscalers and AI loads," Sallee said there have been "several new entrants" in the market.
One of those is Williams Companies $(WMB)$, which provides natural gas pipelines. The company can take the same equipment it uses to power compressors that push natural gas through pipelines to create modules that can provide electricity capacity at a data-center, Sallee said.
"We weren't talking about that three years ago," Landek said about behind-the-meter solutions. "Now it's in the forefront of all these conversations," especially for massive projects exceeding 100 megawatts. In some cases, developers are building excess capacity to export surplus power back to the grid, Landek added.
Big Tech companies may increasingly also pursue alternative energy sources as well. Nuclear power has received a surge of attention, with Meta most recently announcing plans to secure up to 6.6 gigawatts of nuclear energy. However, small modular reactors, the latest innovation in the sector, remain at least five years away from commercial viability according to expert estimates.
While nuclear energy holds much promise for solving energy needs, Landek anticipates that the technology could lead to more controversy once it arrives. "A community needs to be accepting of a data center," he said. "Add a small scale nuclear reactor and the problem just got a little bit more complex."
-Britney Nguyen -Christine Ji
This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
January 17, 2026 09:23 ET (14:23 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments