According to TechCrunch, Nvidia has signed a €1 billion partnership with Deutsche Telekom to establish an “AI factory” in Munich that will boost Germany’s AI computing power by 50%. The project, called the “Industrial AI Cloud,” will use more than 1,000 Nvidia DGX B200 systems and RTX Pro Servers featuring up to 10,000 Blackwell GPUs. Early partners include Agile Robots, which will use its bots to install server racks, and Perplexity, which will provide “in-country” AI inferencing to German users. SAP will contribute its Business Technology platform and applications while Deutsche Telekom handles the physical infrastructure. The facility is scheduled to begin operations in early 2026 and is separate from the EU’s AI gigafactory initiative despite similar timing.
Sponsored content — provided for informational and promotional purposes.
The sovereignty angle
Here’s the thing about this deal – it’s hitting all the right political notes for Europe. Deutsche Telekom CEO Tim Höttges basically said it straight up: “Mechanical engineering and industry have made this country strong. But here, too, we are challenged.” They’re positioning this as a sovereignty play, which is smart given Europe’s growing anxiety about depending on US and Chinese tech giants. The facility will comply with German data sovereignty laws, meaning German companies can use powerful AI without their data leaving the country. That’s a huge selling point in today’s regulatory environment.
The technical scale
Let’s talk about those 10,000 Blackwell GPUs for a second. That’s not just impressive – it’s massive computing firepower concentrated in one location. Nvidia‘s DGX B200 systems are their latest generation, and combining them with RTX Pro Servers creates a hybrid setup that can handle both massive AI training workloads and inference at scale. The mention of digital twins and physics-based simulation suggests they’re targeting industrial applications specifically – think automotive, manufacturing, engineering. These aren’t consumer chatbots; this is serious industrial AI that requires serious computational muscle.
The European context
Now, the timing here is fascinating. The EU just committed €200 billion to build “AI gigafactories” across the continent, but this project is separate from that initiative. Which makes you wonder – is Europe’s public funding actually enough to compete? Because while €200 billion sounds impressive, it’s spread across the entire EU. Meanwhile, single companies like Nvidia, Microsoft, and Google are individually spending hundreds of billions on AI infrastructure. This Deutsche Telekom partnership shows that even with EU initiatives, private investment is still driving the real action. And honestly, that might be the only way Europe can hope to keep pace.
What this means going forward
So where does this leave European AI sovereignty? Well, it’s a step in the right direction, but there’s a catch. They’re still using Nvidia hardware and software – it’s not exactly homegrown technology. The partnership does give Deutsche Telekom control over the infrastructure and data governance, which matters. But the core AI acceleration still comes from California. Still, having this level of computing power physically located in Germany, managed by a European company, and compliant with local laws? That’s significant progress. If this model works, we’ll probably see more telcos striking similar deals across Europe. The race for sovereign AI infrastructure is just getting started.
