One of the most powerful images from President Donald Trump’s inauguration was the seating arrangement: a second row typically reserved for diplomatic dignitaries was instead filled with America’s wealthiest tech titans — Meta’s Zuckerberg, Amazon’s Bezos, Google’s Pichai, and Trump confidant Elon Musk among them.
The visual was a stark departure from tradition, but a clear signal of where influence lies in Trump’s second term. These tech executives, many of whom had previously aligned with Democrats, have pivoted sharply toward Trump. And they’re all circling around the same prize: America’s next great infrastructure boom in artificial intelligence.
The AI-energy nexus, because of its potent mix of intense competition, cash-rich players, and culture of innovation, presents an opportunity to keep the tide rising for the zero-carbon market. The question, though, is whether the president’s vindictive rejection of that market will end up hurting the very tech power players he has chosen to make his allies.
The administration wasted no time setting his AI energy infrastructure agenda. Within hours of taking office, Trump rolled back Biden-era guardrails on AI development, as well as his climate policies.
Energy and artificial intelligence, inextricably linked, appeared throughout the executive orders signed on Inauguration Day, either by name or implication. And even before he took office — In confirmation hearings, as talking points on Sunday talk shows, and in countless hours of discourse online and in podcasts — the need to power the artificial intelligence revolution was cast as a matter of national security and strategic importance.
By Tuesday afternoon, Trump was in the Roosevelt Room announcing “Stargate” — a $100 billion joint venture between OpenAI, Oracle, and SoftBank to finance and build the world’s largest data center supercomputers. The initiative could expand to $500 billion over four years.
“We’re going to make it as easy as it can be,” declared Trump, promising emergency declarations to help these facilities generate their own power. OpenAI’s Sam Altman, fresh from the inauguration, joined him for the announcement.
This rise in tech industry power mirrors the parallel downturn in the clean energy industry’s influence. And while the inversion isn’t a surprise, it’s one that actually could be counter-productive to the task of powering the AI boom.
The clean energy industry is understandably anxious. Significant changes are coming for EVs, offshore wind and climate-friendly regulations, as the first round of executive orders made clear. The incoherence of declaring an energy emergency in order to speed fossil fuel development while at the same time hampering renewables deployment is a familiar, infuriating, Trump posture. But the focus on AI combined with tech companies’ commitments to emissions reductions keeps buyers in the market who are willing to pay a premium for low-carbon energy.
The dominant players — major tech companies, utilities, integrated energy companies, and data center developers — are enjoying historically high free cash flow and strong balance sheets.
The demand they are collectively creating (whether you think it’s going to be an extra 25 GW or 50 GW by 2030 in the U.S. makes little difference) quickly surpassed any grid planners forecast, and certainly surpassed any ability to power it with only fossil fuels. The interconnection queue is almost all renewables and battery storage today. And recent moves by the hyperscalers to support tariffs that speed development of clean firm energy like geothermal remain positive signals — and don’t appear to run afoul of the Trump administration’s favored solutions.
What we know so far
But it’s still early, and the AI-energy nexus is evolving by the day. That said, a few things are clear, even on week one of the Trump presidency.
First, the oil majors will be taking part in powering AI. Their preference will be gas plants, and the easiest play here is to build those plants for major data center operators and provide them as a behind-the-meter, colocated solution.
While potentially disastrous for the country’s emissions, a silver lining may be how this activity gets the carbon capture and storage (CCS) market really humming. Tech companies will likely prefer this approach, at least in the near term, and the addition of CCS can allow it to access further support from government subsidies. This on-site powering approach may also allow utilities and system operators some breathing room as these new data centers won’t be asking for grid power.
The oil majors also have the ability to build these almost anywhere, so this model can further expand secondary and tertiary markets for data centers in ways that wind, solar and nuclear cannot. Executive orders promising easier permitting add an obvious tailwind here.
And the data centers are going to be big. There has been a lot of discussion around the optimal size for data centers as they face land, water, connectivity, and energy constraints. This administration loves the aura of major announcements and AI presents a golden one. By 5:30 pm on Tuesday, Trump had announced in a White House briefing that OpenAI, Oracle and SoftBank had pledged to form a joint venture called Stargate, which would commit up to $500 billion over the next four years. Larry Ellison, Oracle’s chief, said its data center in Abilene, TX, was already under construction via a deal with Crusoe Energy, with many more planned.
With some back-of-the-envelope math — assuming all-in costs for a data center are $25 billion per GW — this represents a commitment of at least 20 GW of data center capacity.
So evidently, the hyperscalers are committed, and they can afford it. Google, Amazon, Meta, Microsoft, OpenAI and other major tech players building data centers are all quite competitive with each other, and with China. And unlike some of the players in past infrastructure bubbles, these companies all are immensely profitable today, have significant free cash flow, and have few regulatory constraints.
And starting this week, they have an administration that has loudly announced its support for their mission and rescinded the one Biden-era executive order that called for the “safe, secure, and trustworthy development and use of artificial intelligence.” No guardrails, fewer permitting hurdles, and the gloss of national security and economic development is casting quite a halo over these players as the year begins.
BCG, for its part, believes these companies are on track to devote $1.3 trillion in capital globally to data centers over the next 5 years.
A version of this story was published in the AI-Energy Nexus newsletter on January 22. Subscribe to get pieces like this — plus expert analysis, original reporting, and curated resources — in your inbox every Wednesday.
