The U.S. is leading the world race when it comes to massive, gigawatt-scale AI infrastructure developments.
The Stargate project alone, with its target of $500 billion in investment and 15 gigawatts in power, would be enough to put the country in the lead, without counting Meta’s $10-billion, 2-gigawatt-plus project in Louisiana. Both projects will initially rely on natural gas, which at least in the short term seems to be inescapable as hyperscalers rush to secure scarce electricity supplies.
But AI, and the question of powering it, is a global challenge. Multi-billion dollar investment pledges in AI infrastructure and promises of massive gigawatt-scale data centers have extended well beyond the U.S. — as have early warnings about a bubble forming.
Europe
In February, the European Union finally formalized its own AI ambitions with the InvestAI initiative, which aims to mobilize €200 billion for AI investments, including €20 billion for AI gigafactories.
So far, one of the most notable European projects in terms of size and money is a 1-GW data center that will be built in France, with an investment of between €30 billion and €50 billion by France and the United Arab Emirates. (“Oh look, a mini Stargate, how quaint,” quipped the technology news site The Register.) While France still has to disclose details about the project, President Emmanuel Macron is betting on his country’s large nuclear capabilities to power its AI developments. “I have a good friend [on the] other side of the ocean, he says drill, baby, drill,’” Macron said at the AI Action Summit in Paris in February. “Here there is no need to drill, it is plug, baby, plug.”
And in Portugal, Start Campus has secured power for a 1.2-GW data center which is set to become the largest in Europe and claims it will use “100% renewable energy.”
The Middle East
While the U.S., China, and Europe have grabbed headlines, many others are also putting some weight behind AI development.
Several countries in the Middle East have been particularly aggressive and ambitious about turning the region into a global AI infrastructure hub. For instance, before there was Project Stargate there was Project Transcendence, a Saudi initiative with as much as $100 billion in backing to invest in AI data centers, infrastructure, and startups announced in late 2024. Also in Saudi Arabia, the NEOM project is a 1.5-GW AI data center which is designed to operate at net zero and leverage both renewable energy and green hydrogen.
And the United Arab Emirates tech fund MGX is a major investor both in Stargate and France’s 1-GW AI data center. And ADQ, an UAE’s sovereign investor, recently announced a $25 billion partnership to invest in power generation for data centers, mostly in the U.S.
East Asia
In China, developers have announced over 500 new data center projects between 2023 and 2024, and major private companies like Alibaba Group and ByteDance have plans to invest $50 billion and $20 billion respectively in AI data centers.
And South Korea explicitly aims to become one of the top global AI powers. The country’s federal government pledged an investment of $7 billion in AI R&D through 2027, but the private sector is making dramatic commitments of its own.
In February, the investment firm Fir Hills, co-founded by Brian Koo of the LG Group’s founding family, announced the development of a 3-GW data center in the Jeollanam-do province of South Korea. The project is estimated to require from $10 to $35 billion in investment — and if it succeeds, it would become one of the biggest data centers in the world. The company hasn’t yet announced official plans for how it plans to power the complex, but has hinted at using renewable energy.
Additionally, SK Telecom’s AI Infrastructure Superhighway program, which was announced in November 2024, plans to develop multiple gigawatt-scale AI data centers AIDCs across Korea.
Lately, however, there have been signs that China may have overestimated demand. Many of the country’s new data centers are reportedly sitting empty and unused — and last week Alibaba Group chairman Joe Tsai warned of a potential bubble forming, noting that developers are raising millions and millions before securing customer contracts.
“Many of the largest technology companies in the world are making investments that kind of only make sense if you absolutely believe the hype about AI in its most extreme forms,” Gregory Allen, director of the Wadhwani AI Center at the Center for Strategic & International Studies, said in a February panel on South Korea’s AI ambitions.
Bubble concerns are beginning to crop up outside of China too — the most notable recent one in the U.S. being CoreWeave’s disappointing IPO last week.
The IPO was influenced in part by the news that Microsoft was pulling back from data center leases because of oversupply issues. And it’s not just the company’s core market of the U.S. that’s impacted; Microsoft had committed to investing $80 billion in AI data centers in 2025 alone, but has now canceled or delayed development plans all over the world, from North Dakota to Australia to Indonesia.
A version of this story was published in the AI-Energy Nexus newsletter on April 2. Subscribe to get pieces like this — plus expert analysis, original reporting, and curated resources — in your inbox every Wednesday.
