Nscale

Nscale has raised the largest Series C round in Europe at $2 billion, led by Aker ASA and 8090 Industries. This round values Nscale at $14.6 billion. This investment fuels our global expansion, accelerates regional capacity, expands our engineering and operations teams, and strengthens the platform layer, powering training and inference at scale. Today, we also welcome three new Directors to our board: Sheryl Sandberg, Susan Decker, and Nick Clegg. Each brings substantial global depth across technology, policy, operations, and governance to an already world-class collection of business leaders. Hear from our Founder & CEO Josh Payne on what’s next. Read the full press release here: nscale.com/press-releases…

Read the full article: nscale.com/blog/the-shift…

The inference ramp-up is here. And most infrastructure wasn't built for it. AI-native infrastructure changes the equation: no legacy constraints, no general-purpose overhead, every layer optimized for AI workloads from day one. Where your compute sits and how your stack is designed is now a strategic decision, not an operational one. Hamish Jackson-Mee, our VP of Product and Design, on what that means for enterprise AI and how we're building for it.

The next chapter in AI will be defined by energy as much as models. At #DellTechWorld, Nscale CRO, Tom Burke, will highlight what functional, scalable AI infrastructure looks like when efficiency and sovereignty are viewed as a design choice, not a trade-off. Find us there: nscale.com/events/dell-te… 📅 Tuesday 19 May | 12:00–1:00 PM 📍Room 4 – Palazzo M, The Venetian

The speed-vs-control trade-off in AI services isn't inevitable. It's a design problem. We wrote about how to fix it: modular infrastructure, built-in governance, and inference optimized at the unit cost level. Read the full article: nscale.com/blog/ai-servic…

Nscale and @BTGroup have announced plans to deliver sovereign AI data centers across three existing BT sites in the UK, powered by @nvidia. Read more here: nscale.com/press-releases…

AI infrastructure is now a board-level topic across banking, government, and enterprise. The physics changed. The financing changed. The operating model has to change too. Nscale's Nidhi Chappell tells strategy+business what C-suite leaders are still underestimating. Read it: strategy-business.com/article/How-ne…

AI spend is no longer a fixed line item. Every interaction is measured in tokens, and levers like applications, compute, storage, networking, and power determine what each token ultimately costs. Nscale on NVIDIA Blackwell optimizes at every level. The result: ~35x lower cost per million tokens than previous generations. That's token economics done right.

We're heading to #DellTechWorld 2026 on May 18. 📍Visit us at Booth #1100 in Las Vegas, where we'll be joining the enterprise AI community to explore the future of AI infrastructure. Nscale will be showcasing our GPU cloud platform for enterprise AI: sovereign, sustainable, and production-ready at scale. Book time with the Nscale team at #DellTechWorld: nscale.com/events/dell-te…

Nscale is expanding its agreement with Microsoft in Norway. More than 30,000 NVIDIA Rubin GPUs will be deployed at our 230 MW Narvik campus in 2027, making it one of the largest onshore AI infrastructure projects in the country. nscale.com/press-releases…

Explore the Kimi K2.5 agentic model and start building on Nscale: nscale.com/blog/kimi-k2-5

Kimi K2.5 can be deployed within minutes on Nscale, meaning you can experiment with Nscale’s Prompt Workbench, get access to a fully-managed production inference, and auto-scale from prototype to production. Nscale uses low-latency and high-throughput batch support, and strict customer isolation, which allows developers to run demanding AI workloads faster, at scale, with enterprise-grade security and reliability.

Kimi K2.5 is now available on Nscale. @Kimi_Moonshot Kimi K2.5 is a 1T-parameter multimodal, open-weight model that transforms text and images into production-ready code and powers agentic workflows, ideal for AI-natives building visual-to-code applications to autonomous multi-agent systems.

Nscale will be the first European company to deploy NVIDIA Vera Rubin at scale. Announced at GTC in March, we’re bringing more than 100,000 GPUs online in early 2027 across the UK, Norway, and beyond. This means European innovators will get early access to some of the most advanced AI infrastructure in the world. For more, read nscale.com/press-releases…

Nscale has partnered with @Fortum to explore plans to build a new data center in Harjavalta, Finland, with Fortum serving as the site development partner. The project aligns with Nscale’s commitment to delivering cutting-edge, sustainable, high-density compute capacity in Europe, expanding our global footprint. Learn more: nscale.com/press-releases…

AI roaming is one of the defining capabilities of distributed AI infrastructure. In the same way mobile networks hand off connectivity as you move between countries, AI roaming enables live sessions and memory to follow users seamlessly across environments. Always-on inferencing is going to require an enriched service. As that becomes the standard, roaming becomes a requirement. nscale.com/blog/ai-roamin…

AI is moving closer to the real world. Real-time workloads won’t tolerate distance, which puts more focus on compute closer to the edge. Telcos are already there. Learn more: nscale.com/resources/repo…

Earlier this month, we announced a landmark collaboration with @Microsoft and @CaterpillarInc to build advanced data center infrastructure for a large-scale deployment of @/NVIDIA Vera Rubin GPUs. Starting in 2027, this will create one of the largest dedicated AI compute installations in the world. nscale.com/press-releases…

Predictable latency is critical for modern AI. We spoke with Marika Mentula, VP of Network Infrastructure at @nokia, about why traditional networks struggle to meet the demands of AI workloads and why consistent, low-latency performance matters more than ever. Watch the full interview: nscale.com/blog/ai-grid-i…