Eindhoven University of Technology strengthens its position as a leading European tech university with a sovereign, secure and future-ready AI facility.
Strengthening TU/e’s leadership in European tech
Eindhoven University of Technology (TU/e) has one clear goal. To continue advancing European technology and research, and keep attracting brilliant minds from around the world.
To enable that ambition, TU/e needed an AI infrastructure it could fully own, control and rely on.
Public cloud environments offer flexibility, but not the control, sovereignty or predictability TU/e needed. Costs fluctuate, performance depends on shared resources, and data control is limited. For sensitive research workloads, the university wanted the opposite: consistent performance, full ownership of its data and a setup with no surprises.
Scaling research without losing control
AI research at TU/e was growing fast. The workloads got heavier, the data sets larger, and the regulations tighter.
In many research environments, shared cloud resources introduce wait times and unpredictable costs. For TU/e, the priority was to remove this uncertainty entirely by owning the full AI stack.
So the mission was clear. Accelerate compute speed, keep every byte of data under TU/e ownership, protect it from end to end, and do it sustainably.
Why TU/e chose to own its AI future
TU/e selected MDCS.AI after a European tender process to design and implement a sovereign AI infrastructure together with NVIDIA and other technology partners like DDN and Borealis.
“Our role was to integrate every layer of the stack, from hardware to software to lifecycle support, so everything works together as one,” said Niels van Rees, co-founder at MDCS.AI.
The aim was simple to describe, but bold to deliver. TU/e set out to build one of Europe’s most advanced high-performance AI facilities owned by the university, powered by sustainable energy, and ready to scale with every new generation of models.
Designing a sovereign, secure, sustainable AI facility
At the centre of the new setup are NVIDIA DGX B200 systems, known for the highest performance per watt. Each system combines powerful GPUs, high-speed networking, and DDN storage for fast, stable, high-throughout data flow.
All of it runs in a Finland data centre chosen for its cold climate and green energy. The facility runs entirely on renewable energy sources, cutting CO₂ output while keeping systems cool and efficient.
Lifecycle management gives TU/e the flexibility to upgrade to newer GPU generations as they become available, or continue with the current setup. This keeps performance high and costs predictable.
Security is built in from the start, with strict access controls, and all data kept within the EU.
As Richard Zoontjes, Lead of the TU/e Supercomputing Center, explains, the ability to compute at scale is essential for competitive research. The new facility gives TU/e researchers the freedom to innovate without limits, with full control, security and sovereignty.
Faster research. Lower risk. Clear ownership.
Here’s what TU/e now experiences every day:
- Predictable performance: researchers have dedicated compute access when needed, without relying on shared cloud queues.
- Full sovereignty: all research data and models stay inside TU/e’s own infrastructure, under complete control.
- True sustainability: the Finland facility runs entirely on renewable energy and efficient heat exchange, cutting the footprint while keeping GPUs cool and efficient.
- Simplicity in operations: MDCS.AI manages the AI Infrastrcuture from design to lifecycle, giving TU/e one accountable partner for the entire stack.
The impact is immediate: with the newest GPUs, high-speed networking, and tuned storage working as one, researchers can train larger models at speeds they have never experienced before.
Early feedback shows significantly faster training times and far more predictable throughput.
A decade of discovery powered by sovereign AI
With its new facility in place, TU/e is ready for a new wave of discovery in generative AI, robotics, and data-driven engineering.
This isn’t just a hardware upgrade. It’s a shift in how science happens, giving researchers freedom to explore without limits and giving TU/e a platform that attracts the brightest new talent.
For MDCS.AI, this case proves its promise: Your AI. Your way. Built on a secure, sovereign, and scalable foundation.
Watch the story
See how TU/e built its AI advantage
Results snapshot
| Objective | Outcome |
|---|---|
| High-performance AI compute | NVIDIA DGX B200 cluster delivers top performance per watt |
| Data sovereignty & security | 100 % EU-based operation with confidential computing enabled |
| Sustainability | Finland data centre powered entirely by renewables |
| Predictability & cost control | Lifecycle management keeps budgets stable |
What’s next
By bringing compute home, TU/e has turned infrastructure into a real advantage. Research runs faster. Energy use is lower. Talent wants to be here. It’s a blueprint other European institutions can follow, one that is secure, sovereign and performance-driven. TU/e owns its AI future. Now it can focus on shaping everyone else’s.
Want to see what sovereign AI can do for your organization?
Contact mdcs.ai to start your AI journey, your way.
