The race to deploy sovereign AI has shifted from a sprint to a marathon of infrastructure endurance. Securing GPUs is only half the battle. The real challenge lies in taming the stack’s complexity, bridging the gap between high-performance liquid cooling and sophisticated cloud orchestration.
With its acquisition of Radian Arc, Barcelona-based Submer has closed the loop, transforming from a liquid-cooling pioneer into a powerhouse for end-to-end, full-stack AI infrastructure.
From Cooling the Chip to Controlling the Cloud
For a decade, Submer has been synonymous with the physical evolution of the data center, leading the charge in immersion cooling. However, today’s announcement signals a strategic pivot. By integrating Radian Arc’s Infrastructure-as-a-Service (IaaS) platform, Submer is moving up the stack.
The acquisition melds Radian Arc’s carrier-embedded GPU edge platform with InferX, Submer’s NVIDIA Cloud Partner (NCP) platform. The result? A dual-plane offering that addresses the two biggest pain points in the current market: latency and sovereignty.
Why the Edge is the New Front Line
While hyperscalers dominate the central cloud, the next frontier of AI, real-time cloud gaming, autonomous systems, and localized LLMs, requires proximity. Radian Arc brings a substantial, pre-existing footprint to the table:
- Global reach with operations across 70+ telecom and edge customers.
- Proven scale with thousands of GPUs already live and are monetized.
- Infrastructure within the carrier network enables data processing that never leaves national borders.
“This acquisition of Radian Arc completes our full-stack cloud infrastructure,” said Patrick Smets, CEO at Submer. “Bringing Radian Arc together with InferX, our AI operations and delivery platform, forms a dual-plane, sovereign, telco-focused cloud offering that is highly competitive in today’s AI datacenter market.”
Building the AI Factory of 2026
The synergy between these two companies creates a single point of accountability for enterprises that are tired of juggling multiple vendors. Submer’s full stack now incorporates:
- A complete AI cloud business unit: Combining InferX (NCP), Radian Arc’s sovereign and telco-grade edge cloud capabilities, and AI inference platforms, enabling scalable and monetisable AI workloads through partner-enabled ecosystems
- End-to-end design and build capabilities: Delivering turnkey, modular AI datacenters and supporting large-scale enterprise and hyperscale-class deployments across Europe and the United States
- Deep IT and liquid cooling expertise: Spanning in-house system design, installation, advanced liquid cooling and integrated AI compute platforms with networking and storage, delivered directly or through strategic partners
- Access to significant land and power pipelines: Exceeding 5GW across the UK, USA, India, and the Middle East through partner consortiums, enabling rapid deployment of next-generation AI infrastructure
A Unified Vision for Sovereign AI
For David Cook, CEO of Radian Arc, the move is about acceleration. By joining Submer’s ecosystem, Radian Arc can move faster to meet the demand of operators who need to monetize AI workloads today, not years from now.
“We have built our platform in close cooperation with our customers and partners, allowing us to develop a powerful model that demonstrably works at scale,” Cook said. “By joining Submer’s established partner ecosystem, we are now in a position to accelerate delivery of sovereign AI infrastructure faster and with lower latency to telecoms operators worldwide.”
As the industry moves away from generic cloud computing toward specialized AI factories, the Submer-Radian Arc union provides a blueprint for what a modern provider looks like: Liquid-cooled, edge-distributed, and entirely sovereign.
For more information on Submer or Radian Arc, visit submer.com.
