Nscale’s $790M Financing Marks a Shift to Utility-Style Deals

extract

Nscale’s $790M Financing Marks a Shift to Utility-Style Deals A $790 million financing package for Nscale’s Norway campus highlights how AI infrastructure is increasingly competing for energy capacity, industrial capital, and grid access. Nordic and European lenders are pouring another $790 million into Norwegian AI infrastructure developer Nscale. The move underscores how banks increasingly view large GPU campuses as long-duration industrial infrastructure tied to energy access rather than speculative tech expansion. The financing package, backed by ABN A, DNB Bank, Eksfin, Nordea, and SEB, will fund continued development of Nscale’s AI data center campus in Narvik, Norway, which Nscale describes as the country’s largest AI infrastructure investment. The agreement also includes a $790 million accordion facility tied to a potential 115 MW expansion, a financing structure more commonly used in utility and industrial infrastructure projects than traditional technology deals. Nordic Power Markets Become AI Battleground The financing highlights how the global AI infrastructure race is shifting from a pure semiconductor competition into a scramble for electricity, land, and capital. Across the Nordics, developers, utilities, and lenders are backing large-scale GPU campuses as hyperscalers and AI firms hunt for regions with abundant renewable energy, available grid capacity, and colder climates that can lower cooling costs for power-dense AI systems. Industry analysts and operators across the region project significant growth in Nordic data center capacity over the next decade as AI infrastructure demand reshapes European power markets. Norway’s hydropower resources and cooler climate have made the region increasingly attractive to operators seeking to contain the soaring operating costs associated with AI training and inference clusters. Yet the surge in AI development is beginning to pressure regional grids. Denmark’s grid operator, Energinet, recently said it had received roughly 60 GW of new grid connection requests, including approximately 14 GW tied to data centers – despite the country’s peak electricity demand sitting near 7 GW. Narvik, located above the Arctic Circle, is part of a broader push across Nordic markets to attract hyperscale and AI computing deployments as operators race to secure long-term power for next-generation GPU clusters. “It’s a rational response to a power problem,” said Sanjay Gupta, co-founder and president of Velaura AI. “When the binding constraint is energy availability, cost, and reliability, you go where the fundamentals are best. The Nordics aren't becoming an AI hub by accident. They're becoming one because power availability is increasingly the deciding factor in where infrastructure gets built.” Nscale Pushes Further into the Neocloud Market The financing builds on a string of recent capital raises by Nscale, including a $2 billion Series C funding round announced in March and a $1.4 billion delayed draw term loan secured earlier this year. Nscale is part of a growing class of “neocloud” providers – GPU-centric infrastructure companies positioned as alternatives to traditional hyperscalers – aiming to offer dedicated AI compute capacity optimized for training and inference workloads. Unlike general-purpose cloud providers, neocloud operators typically concentrate on high-density GPU deployments, faster infrastructure rollout cycles, and direct access to Nvidia hardware for AI-native startups and enterprises struggling to secure capacity through larger public cloud platforms. Nscale positions itself as a vertically integrated AI infrastructure provider spanning energy, data centers, GPU compute, and software services. “Together, these developments position Nscale at the forefront of global AI infrastructure, delivering scalable, high-performance capacity to meet rapidly growing demand for our services,” Nscale founder and CEO Josh Payne said in a statement. AI Financing Starts to Mirror Energy Infrastructure The bank group's commitment also signals growing willingness among traditional lenders to evaluate AI campuses like energy, telecom, or industrial infrastructure projects. “These institutions understand infrastructure cycles,” Gupta said. “They're not underwriting AI hype; they're underwriting power contracts and anchor tenants.” As the cost per inference falls, entirely new applications become viable. “We saw this with cloud computing: more efficiency meant more usage, not less," he said.