America wants to win the AI race.
America also wants cheaper electricity bills.
And right now, those two goals are starting to fight each other.
Wedbush tech analyst Dan Ives is warning that President Trump’s new focus on preventing households from “picking up the tab” for Big Tech’s data center boom could end up creating a new bottleneck: slowing down hyperscaler buildouts at the exact moment the US needs to be accelerating them.
Trump Tells Big Tech to “Pay Their Own Way”
Trump has been pushing the message that the AI data center boom must not raise electricity prices for American households, calling on tech giants to absorb those costs instead.
Microsoft $MSFT ( ▼ 1.85% ) already responded with a “community-first AI infrastructure plan,” committing to cover the electricity costs its data centers create and supporting pricing models where “Very Large Customers” like hyperscalers pay more.
Ives expects other Big Tech companies to roll out similar plans soon.
Ives: This Could Become a Major Data Center Bottleneck
Ives argues that while these plans may reduce political backlash, they could also slow down the pace of AI infrastructure expansion.
His concern is that forcing hyperscalers to internalize these energy costs could:
pressure margins
increase friction with utilities and regulators
delay projects
ultimately slow down data center buildouts
And this comes as the US is entering what Ives calls a “crucial time” in the AI revolution, with growing energy shortages and grid constraints.
China Has the Advantage on Power and Regulation
The competitive urgency here is not theoretical.
In November, Nvidia $NVDA ( ▼ 2.05% ) CEO Jensen Huang said China is going to win the AI race because it has a more favorable regulatory environment and cheaper access to power.
Ives echoed that theme, pointing to China spending more on new and existing power technologies through 2030, increasing pressure on the US to fuel its AI ambitions.
Not All Hyperscalers Will Feel This Equally
In general, hyperscalers have enough profitability to absorb higher costs.
Despite rising depreciation expenses from massive AI capex, estimated profit margins across the hyperscaler group have continued trending higher and remain well above the S&P 500.
But there is one standout laggard: Meta $META ( ▼ 2.32% ).
Meta is more exposed because it does not have a cloud business to directly monetize AI infrastructure. Its payoff is more downstream, largely through advertising, which makes cost absorption harder to justify in the near term.
Meta has also been the only hyperscaler whose projected profit margin has fallen since the end of 2024.
Bottom line: the US is trying to manage the political backlash from AI-driven power demand while simultaneously racing to build AI infrastructure fast enough to stay ahead of China. Dan Ives thinks Trump’s push to protect household electricity bills may be well-intentioned, but could slow down data center buildouts at the worst possible moment.