Scaleway Opens Nvidia-Powered AI Site in Copenhagen, Expanding Sovereign Cloud Capacity

French cloud provider Scaleway announced on Tuesday it will open a new AI computing site in Copenhagen running on Nvidia hardware, its second northern Europe expansion disclosed the same day. Earlier, the company said it would build an AMD-powered supercomputer in Sweden, a twin move aimed squarely at European organizations that want their AI work hosted under EU jurisdiction (Reuters, 2026).

The Copenhagen facility will use Nvidia's HGX B300 system built around the chipmaker's latest-generation accelerators. Scaleway said the city was chosen for its technology sector, low-carbon electricity, and cool climate that reduces cooling energy. The company already serves Danish customers including the City of Copenhagen and AI firm Corti (Reuters, 2026).

CEO Damien Lucas framed the expansion around a simple premise. "Access to high-performance computing for AI is becoming a strategic priority for European organizations," he said. That priority has both a technical and a regulatory dimension. European businesses in healthcare, finance, and the public sector increasingly need compute capacity that sits inside EU borders, subject to EU data protection rules rather than US or Chinese jurisdiction.

Scaleway's decision to split its northern expansion between Nvidia and AMD silicon is notable. It signals a willingness to offer customers architectural choice rather than locking them into a single accelerator ecosystem. For AI developers evaluating European cloud options, that flexibility could matter as much as raw performance.

The timing places Scaleway in direct competition with larger hyperscalers building their own EU-sovereign capacity, including Microsoft, Google, and a growing roster of European providers. The company's advantage is not scale but focus. It sells a specific proposition: European compute, European law, and now two hardware paths to choose from.

BuiltWorld AI Operational Take: Scaleway's dual-silicon strategy adds operational complexity at the facility level. Supporting both Nvidia and AMD platforms in northern Europe means duplicate supply chains for spare parts, separate power and cooling profiles, and staff trained across two architectures. Whether that flexibility becomes a selling point or a cost burden depends on execution. For project teams planning sovereign AI capacity, the choice is no longer just where to build but how many hardware ecosystems to support.