

AMS1 —Amsterdam’s HPC Facility
Our AI-Ready Compute Suite, AMS1, is a 6 MW vendor-agnostic data hall designed for Bring-Your-Own-Server and racks, that is purpose-built for AI and HPC workloads. The facility supports 150+ kW per rack via dual-fed overhead busway with tap-off boxes and advanced cooling — fully prepared for liquid cooling, featuring an optional pre-design for direct liquid cooling (DCLC) and rear door heat exchangers (RDHx).
6
IT Capacity Available Now
150+ kW
Per Rack
7610
IT Space
Your Server, Rack or AI Pod Design
Use your own rack, server or AI Pod designs — vendor-agnostic and sovereign-by-design. Or get pre-design support.
Power Available Now
Access to a 6 MW suite with dual-fed overhead busway (150+ kW per rack),
immediate RFS state.
AI-Ready With Cooling Flexibility
Equipped with the flexibility to use liquid cooling, and a closed loop chilled water system, all designed with an N+1 configuration to maximize energy efficiency and operational reliability.
Unlock Tomorrow's Data Innovation Today
Ardent Data Centers’ AMS1 is a 6 MW, liquid-to-the-rack facility in Amsterdam built for extreme-density AI and HPC. With advanced liquid cooling and carrier-rich connectivity, AMS1 supports NVIDIA Grace Blackwell NVLink pods (NVL36/NVL720), HGX Blackwell/Hopper servers (B200/H200/H100), as well as AMD Instinct MI3 platforms (MI300X/MI325X/MI35x) and their successors. Deploy tightly coupled training or elastic scale-out inference faster, with lower energy and floor-space overhead, without redesigning your stack.
AI-Ready Compute Suite
Our state-of-the-art 6 MW data hall in Amsterdam is purpose-built to meet the demanding needs of AI applications. Designed with flexibility in mind, it offers a customizable IT space, empowering our customers to create bespoke environments tailored to their unique requirements. Fully optimized for AI workloads, this facility ensures seamless performance and scalability for cutting-edge innovations. Optional fitout design available with advanced cooling packages including DCLC and RDHX to boost efficiency.
Disclaimer: This video is from a similar facility in London. Layout, equipment, and specifications may differ from our Amsterdam site.
Technical Specifications
IT Space sq ft | 7,610 sq ft |
Rack Support | Support up to 52U racks |
Cooling System | Liquid cooling ready for DCLC or RDHx, closed-loop chilled water |
Cooling Reliability | Concurrently maintainable, N+1 design |
IT Capacity Available Now | 6 MW |
Total IT Capacity | 6 MW |
Carrier Neutral | Multiple carrier options |
Meet-Me Rooms | Two available |
Exterior Security | 24 × 7 × 365 on-site security |
Interior Security | Video surveillance, badge access, biometric access controls |
Tier Alignment | Tier III aligned – N+1 power & cooling |
Certifications | ISO 27001 |
Fire Detection | VESDA |
Strategic Benefits for Your Business
Ultra-Low Latency
Gain a critical advantage with direct, low-latency access to global financial centers, trading platforms, and European markets through Amsterdam's digital gateway and cloud adjacent locality.
European Sovereignty
Operate with confidence in a Tier 1 European market with strong compliance standards and robust security protocols for data sovereignty.
Future-Proofed Infrastructure
Scale effortlessly with a facility designed to support evolving technologies and increasing power demands, ensuring your investment is ready for tomorrow's challenges.
.png?width=4258&height=3899&name=ND_MD_Rack_Double%20(1).png)
Optional Fitout
for Peace of Mind
Our reference architecture is designed with scalability at its core, enabling adaptable layouts, power, and cooling solutions to grow with demand. Engineered for optimized efficiency, it ensures seamless integration while balancing high performance with sustainability. Supporting rack densities of over 200 kW, the system maintains peak performance while minimizing energy consumption. As a dedicated service partner, we deliver bespoke design solutions tailored to support NVL36/NVL72, HGX B/H-series, and AMD Instinct MI3xx footprints, empowering next-generation AI and HPC workloads.
