“`html

SoftBank Commits €75 Billion to French Data Centers

SoftBank just pledged up to €75 billion to build data centers across France. That single commitment is larger than the entire GDP of Luxembourg. And most investors are treating it like background noise. That’s a mistake.

Why This Is Happening Now

France has been quietly making itself into the most attractive place in Europe to build AI infrastructure. The French government’s Choose France initiative has pulled in over €15 billion in foreign tech investment pledges since 2024, according to Reuters. The country runs more than 70% of its national electricity grid on nuclear power, according to Électricité de France, which makes it one of the cheapest and most stable power sources on the continent.

SoftBank announced the €75 billion commitment at a press conference alongside French President Emmanuel Macron, according to the Financial Times. The deal covers multiple large data center campuses spread across several French regions. Construction is expected to run through the early 2030s.

Europe is in a race to build AI compute capacity before American and Chinese firms lock in a permanent lead. According to the International Energy Agency, global data center electricity consumption is expected to exceed 1,000 terawatt-hours annually by 2026, more than double the roughly 460 terawatt-hours recorded in 2022. Someone has to build the facilities to power that demand. SoftBank is betting it should be them, in France.

What Everyone Is Getting Wrong

The business press is framing this as a France story. It’s not. It’s a Son story. More specifically, it’s a power story.

Masayoshi Son doesn’t invest in countries. He invests in asymmetric bets with structural advantages. France’s nuclear grid is that advantage. A data center running on French nuclear power operates at a fraction of the cost of one running on a US natural gas grid. When you’re burning hundreds of megawatts of electricity around the clock to run AI compute, that cost difference compounds into billions of dollars over a decade. Son did the math. He liked the answer.

I’ve watched Son make bets that looked insane and turn out correct. The early Alibaba investment, the ARM Holdings acquisition, the Vision Fund bets that eventually recovered. According to Bloomberg, SoftBank reported net income of approximately $5.6 billion in fiscal year 2024 after two years of heavy Vision Fund losses. He’s not a reckless gambler. He’s a patient, high-conviction investor who bets on structural shifts, not quarterly earnings.

The structural shift here is obvious. AI models need compute. Compute needs power. Power in Europe means nuclear, and nuclear means France. Son is building the roads that every AI company in Europe will eventually need to drive on.

Here’s the contrarian read most people won’t say out loud. The US has a chip advantage, a talent advantage, and a capital markets advantage in AI. But it does not have cheap, carbon-free, baseload power at the scale Europe now has. France, with its Soviet era investment in nuclear capacity, has a 50-year head start on building the kind of power grid that AI infrastructure requires. Son sees that. Most people don’t.

Poor mindset: “SoftBank is investing in France. Interesting headline.” Rich mindset: “The most well-connected tech investor in Asia just told me exactly where AI compute will live in 2035. What do I own that sits upstream of that supply chain?”

The supply chain answer includes uranium producers, power grid equipment makers, cooling technology companies, fiber optic providers, and the French industrial property owners sitting on the land where these campuses get built. None of them made the headline. All of them benefit.

For content creators and small business owners, the downstream impact is cheaper AI tools. More compute capacity means lower inference costs, which means the AI tools you use every day get faster and cheaper. If you’re already building with something like InVideo AI to produce video content at speed, expect those tools to keep improving as this European infrastructure comes online. The underlying cost curve is bending hard in your direction.

What I Would Do With This Information

Let me be specific about who this matters to and how.

If you’re an investor, stop watching AI software stocks for a second and look at what they actually run on. The data center infrastructure buildout in Europe is going to require enormous amounts of specialized equipment. Power transformers, cooling systems, and fiber are all in constrained supply right now. According to Goldman Sachs, global data center capital expenditures are projected to exceed $1 trillion cumulatively through 2030. That money flows somewhere concrete. Find out where before the crowd does.

If you’re an entrepreneur, this matters for a different reason. A massive European compute buildout means cheaper AI compute for European startups, which means better tools at lower prices for everyone building products that serve European customers. The window between “infrastructure gets built” and “prices drop for end users” is where smart builders get ahead. Start building now, before the cost curve catches up with the general market.

If you want to start a side business or online presence but feel priced out of quality software, pay attention to where deals exist right now. Platforms like AppSumo regularly run lifetime deals on AI-powered tools from early-stage startups. Those deals disappear fast once the underlying compute costs drop and these companies raise prices to match demand. I’ve seen this cycle play out before. The best time to stack affordable tools is before everyone else figures out where the market is headed.

The one thing I would not do is wait. Infrastructure commitments of this size take years to complete, but the investment signals move markets immediately. Position around a structural shift before the crowd sees it clearly or don’t bother at all.

The Bottom Line

SoftBank just put €75 billion on France. That’s not a press release. That’s a decade-long construction project built on the thesis that AI runs on nuclear power, and whoever owns the pipes gets paid forever. The apps get the attention. The infrastructure gets the money. Son knows which side he wants to be on. The question is whether you figure that out now or read about it later in someone else’s annual report.

Frequently Asked Questions

Why did SoftBank choose France for this massive data center investment?

France offers two structural advantages that most countries can’t match: stable, low-cost nuclear electricity and strong government support through initiatives like Choose France. According to Électricité de France, more than 70% of France’s electricity comes from nuclear power, making it one of the cheapest and most reliable energy sources in Europe for power-hungry facilities like AI data centers.

How much does €75 billion in French data centers actually build?

At current construction costs, €75 billion is enough to fund multiple gigawatt-scale data center campuses across the country. For comparison, most major individual data center projects in the US cost between $1 billion and $5 billion each. SoftBank’s commitment could fund a significant portion of Europe’s total AI compute capacity over the next decade.

What does the SoftBank French data center push mean for AI costs?

More compute capacity in Europe will put downward pressure on AI inference costs for European companies. As supply increases, prices for AI computing time tend to fall. This means cheaper, faster AI tools for businesses and consumers across the continent over the next few years.

Does this investment challenge US dominance in AI infrastructure?

Not immediately. The US still leads in semiconductor manufacturing, AI research talent, and total compute capacity. But Europe is closing the gap faster than expected, and a €75 billion commitment from one of the world’s most connected tech investors signals that the concentration of AI infrastructure is spreading beyond the US West Coast.

Should regular investors pay attention to the SoftBank French data centers announcement?

I think yes, and sooner rather than later. Big infrastructure commitments like this one signal where capital flows for years. The direct beneficiaries include power equipment makers, data center construction firms, cooling technology companies, and fiber providers. These aren’t glamorous sectors, but they have real pricing power when demand is this concentrated and supply is constrained.

“`