Stargate and EU's AI investments

Two weeks ago, Project Stargate, a massive, high-profile joint venture, was announced by US President Donald Trump that aims to revolutionize the United States’ AI infrastructure.

France and the EU have matched this announcement with equivalently large undertakings. I would like to unpack these announcements. What is it? Where does the money come from? And, what’s the spending plan? Plus, I might comment here and there.

Meanwhile, China is not announcing much but focuses on AI.

The further I got in my research the more frustrated I became. 🙊

Big Bets on AI - The US, France, and the EU

Stargate in the US

What is it?

Stargate involves developing expansive data centers and associated energy systems to power AI.

Where does the money come from?

The funding is sourced from private-sector investments by key partners: OpenAI, Oracle, and SoftBank are the primary backers, with additional capital from investment firm MGX and other limited partners. The initial commitment is about US$100 billion, with plans to eventually invest up to US$500 billion over the next four years.

Where should the money be invested?

The investment is for building and expanding AI data centers (with the first major site in Abilene, Texas) and upgrading the necessary energy infrastructure—such as power plants and related facilities—to meet the high energy demands of these centers.

France’s €109 billion AI investment

What is it?

A massive €109 billion investment package to push AI in the coming years was announced by President Emmanuel Macron at the Paris AI Action Summit.

Where does the money come from?

Most of the €109 billion comes from private and foreign investors, with the UAE contributing up to €50 billion, followed by Brookfield (€20 billion), Bpifrance (€10 billion), and others. The French government coordinates and offers incentives, but contributes only a tiny portion directly.

I have comments to governments coordinating the investments further below, after the EU plan. Positively, France is not spending any (taxpayer) money directly.

Where should the money be invested?

Most of it will fund new data centers and supercomputers. The UAE, Brookfield and other contributors will get their pole position: a mega data campus (1 GW capacity) will be built under a French-Emirati alliance; Brookfield will expand data-center giant Data4; Digital Realty and Equinix plan multi-billion-euro expansions. FluidStack is erecting a nuclear-powered supercomputer, and Mistral AI is building Europe’s largest AI cluster.

The French government seems to have chosen Mistral AI as their national AI champion.

The goal is bold: “match or exceed US and Chinese compute power”

Even though these are significant steps, I doubt they can match the pace of the US and China. They are too far ahead and too determined to push even further. Let’s be real; France alone can not match or exceed. (See comparison table below.)

European Union's €200 Billion AI Investment Plan

What is it?

President Ursula von der Leyen announced the plan to mobilize €200 billion for AI for the EU region. This initiative is unprecedented in scale for Europe. The plan, dubbed InvestAI, is combining funding from EU institutions, member states, and industry.

Where does the money come from?

Roughly €150 billion in private-sector commitments and ~€50 billion in public financing via the European Commission and the European Investment Bank (EIB).

While there’s a nice chunk from the private sector, I think the roughly €50 billion from public financing is wrong. Governments are known to be bad at investing money. Pouring this amount into AI will lead to huge waste (as history shows), and this gives socialist vibes. Making it easier to unlock risk capital from real institutions is key to actual progress while reducing regulatory hurdles.

Where should the money be invested?

• Building 4 AI gigafactories with 100k+ top GPUs → for training models

• Venture funding, equity co-investments, and innovation grants to help AI companies grow inside the EU (rather than relocating to the US).

• AI adoption in the industry

• Supporting “Common European Data Spaces” (shared datasets under GDPR rules) and expanding cloud/edge projects (e.g. GAIA-X) to give European AI teams robust local infrastructure.

The EU sees itself as an orchestration layer. I don’t think that works out well.→ remove hurdles and drive growth with the right incentives, but don’t manage the money!

As a government, you shouldn’t interfere(“here is money, make an awesome product” without actually winning in a real market). In fact, you should remove yourself as much as possible so actors can innovate in an internationally competitive environment.

Ground rules must be established, don’t get me wrong. But looking at the current non-well-co-organized regulatory landscape of IT, digital, AI, and tech regulations, we are way past ground rules.

Let’s assume, we want to build AI in healthcare.

For a typical EU AI project, you have to deal with 11 major regulatory frameworks. In detail, these include the AI Act, GDPR, Digital Services Act, Digital Markets Act, Cybersecurity Act, NIS2 Directive, Competition Law, Copyright/IP rules, ethical/transparency guidelines, plus the two healthcare‐specific rules (MDR and IVDR).

Collectively, these frameworks break down into roughly 150–200 distinct sub regulatory provisions you must comply with. (In reality, there is even more.)

If you don’t comply, fines are coming:

• For the EU AI Act, up to €35 million or 7% of global annual turnover (whichever is higher)​

• For GDPR, up to €20 million or 4% of global annual turnover (whichever is higher)​

• For the Digital Service Act, up to 6% of global annual turnover (whichever is higher)​

• For DMA, up to 10% of global annual turnover

• NIS2 Directive, up to €10 million or 2% of global annual turnover (whichever is higher)​

• Competition Law, 10% of global annual turnover

• Copyright infringements, MDR, and IVDR, open ceiling. Product recalls, bans, and potential bankruptcy.

While these regulations all make sense from their point of view, there is no harmonization between them. This is the biggest problem, according to lawyers and data protection officers I talked to at this year’s CISOx conference in Munich.

No harmonization leads to contradicting rules at times. It leads to an unclear regulatory environment. Expert teams need to be hired. And, ultimately, hinders risk capital, innovation, and bold bets.

I listened to Ursula von der Leyen’s speech. While I like the enthusiasm around moving forward in AI within the EU, I think it was delusional. She pointed out that the EU is not behind in AI.

The contrary is the case (Am I too pessimistic? Let me know.):

This is by no means a comprehensive list. Other factors like talent, AI adoption, etc., are also in favor of the US.

Last fact: 16 years ago the US market capitalisation was equal to the one from the EU. Today, it is 3:1.

So, to say that the EU wants to “match or exceed US and Chinese” in anything AI-related is delusional.

We (I am a proud European, and very much enjoy my live here, in Germany) made the first step, but I pray that the EU governments don’t repeat socialistic money management, leading to money wasting rather than wealth generation.

I hope you enjoyed it.

Happy weekend!
Martin 🙇

• I recommend:

  • Beehiiv if you write newsletters.

  • Superhuman if you write a lot of emails.

  • Cursor if you code a lot.

  • Bolt.new for full-stack development.

  • My book - Generative AI: Navigating the Course to AGI.

  • Follow me on X.com.

• AI for your org: We build custom AI solutions for half the market price and time (building with AI Agents). Contact us to know more.

You might like our last episodes:

1  https://geopolitique.eu/en/2025/02/10/financing-infrastructure-for-a-competitive-european-ai/#:~:text=far%20as%20data%20centers%20are,AI%20models%20on%20European%20soil

2  https://geopolitique.eu/en/2025/02/10/financing-infrastructure-for-a-competitive-european-ai/#:~:text=compared%20with%2037,the%20French%20company%20Mistral%20AI

3  https://aiindex.stanford.edu/wp-content/uploads/2024/04/HAI_AI-Index-Report-2024_Chapter1.pdf#:~:text=61.1,1

4  https://huggingface.co/spaces/lmarena-ai/chatbot-arena-leaderboard

5  https://www.reuters.com/technology/artificial-intelligence/ai-cloud-funding-us-europe-israel-hit-79-bln-2024-accel-says-2024-10-16/

6  https://geopolitique.eu/en/2025/02/10/financing-infrastructure-for-a-competitive-european-ai/#:~:text=European%20cloud%20companies%20account%20for,expensive%3A%20some%20estimates%20calculate%20the