Whether you're looking for an angel investor, a growth advisor, or just want to connect — I'm always open to great ideas.
Get in TouchNscale's $2B Series C at a $14.6B valuation, backed by Nvidia, Citadel, and Jane Street, cements its position as Europe's leading AI cloud platform.
AI, startups & growth insights. No spam.
Europe has its first true AI hyperscaler. London-headquartered Nscale announced on March 9, 2026 that it has closed a $2 billion Series C funding round at a post-money valuation of $14.6 billion — a nearly 5× jump from its $3.1 billion Series B valuation just four months ago. The round brings in a striking roster of investors: Nvidia, Citadel, Jane Street, Dell, Nokia, Lenovo, and financial firms Astra Capital and Linden Advisors, led by Norwegian industrial giant Aker and 8090 Industries. Alongside the capital, three heavyweight board members join: Sheryl Sandberg (former COO of Meta), Nick Clegg (former UK Deputy Prime Minister and Meta's global affairs president), and Susan Decker (former president of Yahoo).
This is not just a large funding round. It is a signal that the race for AI infrastructure is now a global affair — and Europe has a serious contender.
Nscale's Series C is remarkable not just for its size, but for its velocity. The company completed a $1.1 billion Series B in November 2025 at a $3.1 billion valuation, followed by a $433 million pre-Series C SAFE announced in October 2025. In February 2026, Nscale secured a $1.4 billion debt facility specifically for GPU purchases. Now, just months later, the Series C has vaulted the company to a $14.6 billion valuation.
That is a $11.5 billion valuation increase in roughly four months.
For context, that rate of value creation puts Nscale in a class with the fastest-growing technology companies in the world. The investment narrative is straightforward: AI compute demand is growing faster than the infrastructure industry can build to meet it, and whoever controls the most efficient, most scalable, and most strategically situated data centers will capture enormous economic value over the next decade.
Lead investor Aker — the Norwegian industrial conglomerate — is dramatically increasing its commitment. Aker's stake in Nscale rises from 9.3% to 23.9% upon close, making it far and away the dominant shareholder. Aker has been a critical partner since the early days, providing not just capital but strategic access to Norway's abundant renewable energy resources and industrial infrastructure. The partnership with Aker is what makes Nscale's Norwegian data centers — including the landmark Stargate Norway project in Narvik — economically feasible at the scale being built.
8090 Industries co-leads the round. The remaining investors span the full breadth of the AI ecosystem: Nvidia for strategic chip supply and ecosystem alignment; Citadel and Jane Street for sophisticated financial backing; Dell, Nokia, and Lenovo for infrastructure partnership angles; and Point72 for institutional conviction in the growth story.
Nscale describes itself as "the hyperscaler engineered for AI" — and that framing is deliberate. The company is not simply renting GPU capacity the way early cloud companies rented CPU cycles. It owns the entire stack, from the land and power infrastructure, to the data center buildings, to the GPU clusters themselves, to the software platform that researchers and enterprises use to train, fine-tune, and deploy AI models.
The product portfolio breaks into four layers:
AI Services — The top of the stack. Nscale provides serverless inference endpoints (supporting models like Llama, DeepSeek, and Qwen), a Prompt Workbench for iterating on model behavior, and serverless fine-tuning. For companies that want to customize a foundation model on proprietary data without managing GPU clusters, this is the entry point.
Platform Services — For teams that need more control, Nscale offers managed Slurm clusters (the standard in high-performance computing for distributed training jobs), Kubernetes services for containerized workloads, and traditional virtual machine and bare metal instances. This layer targets AI labs, research institutions, and engineering-heavy enterprises.
Infrastructure Services — The raw compute substrate: GPU and CPU instances, high-speed networking, and tiered storage. Nscale competes directly with cloud GPU providers here, but with a structural cost advantage that CEO Josh Payne has quantified publicly: "Our cost of production for compute is at least 10% lower than our competitors."
Fleet Operations — An operational layer that is often overlooked but critically differentiating. Nscale's Control Center, Observability platform, and Radar API give large customers visibility and control over their compute fleets. For an AI lab running thousands of GPUs across multiple jobs, the ability to monitor utilization, diagnose failures, and automate scheduling is not a nice-to-have — it is essential for operational efficiency.
Nscale operates or is actively building data centers across five countries: Norway (Glomfjord and Narvik), the United Kingdom (Loughton), the United States (Texas and North Carolina), Portugal (Sines, via Start Campus), and Iceland (Keflavik). Each location was chosen with deliberate strategy: Norway and Iceland for abundant, cheap hydroelectric power; the UK for proximity to European enterprise customers and sovereign compute requirements; Texas for US hyperscale demand and access to US AI labs.
The investment thesis behind Nscale comes down to three structural dynamics that are reshaping global AI.
First, the compute gap is real and widening. The demand for GPU compute from AI labs, enterprises, and government programs has materially outpaced supply for three consecutive years. Every major AI model release — from GPT-5 to Llama 4 to Gemma 3 — requires more training compute than the last. Meanwhile, the time from GPU production to functional data center deployment is measured in months, not days. Nscale is not speculating about future demand; it has supply agreements with OpenAI and Microsoft today.
Second, energy is the binding constraint — and Nscale has solved it differently than US hyperscalers. A 100-megawatt AI data center in the US requires either connecting to an already-strained power grid or waiting years for new generation capacity. Norway's Glomfjord and Narvik facilities, by contrast, are powered by 100% renewable hydropower that has been available for industrial use for decades. Nscale's Stargate Norway project — co-developed with Aker ASA and OpenAI — is targeting 100,000 NVIDIA GPUs and 230 megawatts of initial capacity, with plans to expand to 520 megawatts. This is not a data center; it is an AI gigafactory.
Third, sovereignty matters more every year. European regulators, governments, and large enterprises are increasingly unwilling to store sensitive data or train foundation models exclusively on US-controlled infrastructure. Nscale's European-first positioning, combined with its planned £2.5 billion investment in the UK over three years, makes it the natural partner for sovereign AI initiatives. The company is already embedded in the UK Sovereign Industry Forum alongside Nvidia, and has commitments to deploy 10,000 NVIDIA Blackwell GPUs in the UK by the end of 2026 and hire 100 AI specialists domestically.
Nvidia's participation in this round deserves particular attention. Nvidia is, first and foremost, a chip company. Its business model has historically been to sell GPUs to whoever will buy them, without preference for which cloud or data center deploys them. Strategic equity investments in compute platforms represent a meaningful evolution of that model.
By taking an equity stake in Nscale, Nvidia accomplishes several things simultaneously. It deepens alignment with a customer that has already committed to deploying tens of thousands of Blackwell GPUs. It gains visibility into how those GPUs are being utilized and what product features would serve Nscale's customers better. And it positions Nvidia as a stakeholder in the long-term success of European AI infrastructure — a politically important signal given the EU's focus on technological sovereignty.
This follows a pattern Nvidia has been developing across the AI stack. The company has invested in numerous AI applications companies (ElevenLabs, Cohere, and others) while also ensuring that its hardware is embedded in the most strategically important infrastructure projects globally. Nscale's Stargate Norway already lists Nvidia as a core technology partner for accelerated computing. The Series C investment formalizes that relationship with equity.
For Nscale, Nvidia's participation carries a signal value that transcends the dollar amount of the investment. Having the world's dominant AI chip company as a named investor makes it substantially easier to secure preferential GPU allocation — a critical competitive advantage when GPU supply remains constrained.
The addition of three high-profile board members alongside the funding is unusual and worth unpacking.
Sheryl Sandberg spent 14 years as Meta's COO and built one of the most successful digital advertising platforms in history. Her relevance to Nscale is strategic: she understands how hyperscale infrastructure becomes the backbone of global digital commerce, and she has decades of experience navigating the relationship between technology platforms and their enterprise customers. For Nscale, which needs to expand beyond its current focus on AI labs toward the broader enterprise market, Sandberg's institutional knowledge of enterprise sales and platform economics is directly applicable.
Nick Clegg served as Deputy Prime Minister of the United Kingdom from 2010 to 2015 and subsequently as Meta's President of Global Affairs until early 2026. Clegg's value to Nscale is unambiguous: he has deep relationships across European governments, regulatory bodies, and multilateral institutions. As AI infrastructure becomes a matter of national policy — with governments actively competing to attract data center investment and shape AI governance frameworks — having a former head of government on the board is a genuine strategic asset. Nscale's £2.5 billion UK investment commitment and its participation in the UK's sovereign AI programs benefit directly from Clegg's relationships and credibility.
Susan Decker served as President of Yahoo during one of the most challenging periods in internet history. Her experience navigating a legacy media and technology company through rapid structural change brings a different kind of wisdom to the board: how to manage organizational complexity, partner with competing platforms, and execute at scale during periods of market disruption. For a company growing as fast as Nscale — which went from Series A to $14.6 billion valuation in roughly 15 months — operational governance and organizational scaling are genuine risks that experienced board oversight can help manage.
Collectively, the three appointments signal that Nscale is preparing for the next phase of its corporate life: deeper enterprise relationships, intensive regulatory engagement across multiple jurisdictions, and the organizational complexity that comes with being a multi-billion-dollar infrastructure company with global operations.
CEO Josh Payne has been direct about Nscale's competitive positioning since the Series A. In his words: "Nscale is the hyperscaler engineered for AI, the antithesis to the Neoclouds."
The "Neocloud" category — GPU rental companies that effectively act as middlemen between Nvidia and AI customers — has grown rapidly but faces inherent structural limitations. Neoclouds do not own their power infrastructure, cannot negotiate the same long-term GPU supply terms as integrated players, and have limited ability to optimize the full stack for AI workloads. Their margins are thin and their competitive moat is shallow.
Nscale's vertical integration — owning land, power contracts, buildings, GPU clusters, and software — creates cost advantages that compound over time. Payne has stated that Nscale's cost of compute production is at least 10% lower than competitors. For a market where customers are spending tens of millions of dollars per year on compute, 10% is material. For customers spending hundreds of millions, it is transformative.
The renewable energy angle reinforces this. Hydropower in Norway is among the cheapest, most reliable electricity in the world. Power costs represent the largest variable operating expense in any data center business. Building on the cheapest power source in Europe is not just an environmental statement — it is a durable cost advantage that US-based competitors cannot easily replicate.
The $1.4 billion debt facility secured in February 2026, structured specifically for GPU purchases, reflects the sophistication of Nscale's financial architecture. Rather than diluting equity to buy hardware, the company is using asset-backed debt — treating GPUs as the collateral they effectively are — to maintain balance sheet efficiency while expanding compute capacity.
Nscale is not without formidable competition. The AI infrastructure space has attracted some of the most well-capitalized companies in the world.
AWS, Azure, and Google Cloud remain the default choices for most enterprise AI workloads. They have the brand recognition, the existing relationships, the compliance certifications, and the breadth of services that Nscale simply cannot match today. However, the hyperscalers face their own constraints: massive existing infrastructure commitments, internal politics around GPU allocation, and — particularly for Azure and AWS — data sovereignty concerns that make European customers hesitant to commit entirely.
CoreWeave, the US-based GPU cloud that went public in early 2025, is the most direct analog to Nscale in the American market. CoreWeave raised billions, focused exclusively on AI compute, and built deep relationships with AI labs. Nscale's European positioning and renewable energy advantage differentiate it from CoreWeave, but the two companies are competing for the same global AI lab and enterprise customers.
The Neocloud tier — companies like Lambda Labs, Coreweave's smaller competitors, and dozens of regional GPU rental services — is where Nscale's vertical integration and scale most decisively differentiate. As Payne has noted, Nscale is positioned as the antithesis to this category.
What Nscale has that few competitors can claim: a combination of sovereign European positioning, renewable-powered data centers at gigafactory scale, supply agreements with the world's two most important AI companies (OpenAI and Microsoft), and now a $14.6 billion valuation that gives it the balance sheet credibility to win the largest enterprise deals.
The $2 billion Series C, combined with the $1.4 billion debt facility already secured, gives Nscale approximately $3.4 billion of new capital to deploy in 2026. The priorities are visible in the company's public commitments:
UK expansion is the most specific public commitment: £2.5 billion in the UK over three years, including the 10,000 NVIDIA Blackwell GPU deployment by end of 2026 and the hiring of 100 AI specialists domestically. Nscale's UK facility in Loughton (Essex) is already operational, and the capital will accelerate its expansion into a full sovereign AI compute hub.
Stargate Norway continues to scale. The Narvik facility, developed in partnership with Aker and OpenAI as the first European AI gigafactory under OpenAI's "For Countries" program, is targeting 230 megawatts of initial capacity with 290 megawatts of expansion planned. Getting to 100,000 NVIDIA GPUs at that site requires both capital for hardware procurement and the operational capability to deploy and manage clusters at that scale.
US expansion is actively underway, with facilities in Texas and North Carolina. The US market represents the largest pool of AI compute demand globally, and Nscale's presence there is essential to serving AI labs and enterprise customers that operate across both sides of the Atlantic.
Platform and software investment will continue to differentiate Nscale's cloud offering from pure infrastructure plays. The serverless inference, fine-tuning, and fleet operations capabilities that constitute Nscale's software layer are not free to build — and as the AI stack matures, software differentiation will matter more, not less.
It is worth stepping back to understand what Nscale's $14.6 billion valuation implies about the market it is pursuing.
The global cloud infrastructure market is already measured in hundreds of billions of dollars per year, with AWS alone generating over $100 billion annually. AI-specific compute — training clusters, inference infrastructure, and the full-stack platforms to manage them — is growing faster than any other segment of enterprise technology.
Europe specifically has a structural gap. The continent is home to world-class AI research institutions, some of the world's largest enterprise companies, and a regulatory environment that demands data sovereignty and energy accountability. Yet until recently, there was no European company capable of providing AI infrastructure at the scale that frontier model training requires. Nscale is the first credible answer to that gap.
The geopolitical dimension matters too. Post-2026, no major government wants to depend entirely on foreign-controlled infrastructure for its AI capabilities. The EU AI Act, which entered enforcement in early 2026, accelerates this dynamic by creating compliance requirements that are most cleanly met on infrastructure that operates under European legal jurisdiction. Nscale's data centers in Norway, the UK, Portugal, and Iceland are not just commercially valuable — they are strategically necessary for European AI sovereignty.
The comparison to the early hyperscaler era is instructive but imperfect. AWS, Azure, and Google Cloud became dominant partly because they got there first and built network effects around developer tooling and data. AI infrastructure is building those same network effects now, but with an added dimension: the compute infrastructure is the bottleneck for capability development itself, not just deployment. Whoever controls the most efficient GPU clusters in the most strategically important locations will shape which AI models get built and which applications get deployed.
Nscale, with its $2 billion Series C, its Nvidia backing, its OpenAI and Microsoft supply agreements, and its renewable-powered gigafactory in the Norwegian Arctic, is betting that it can be that company for Europe — and increasingly for the world.
This is a story worth watching closely. The AI infrastructure race is far from over, and the prize for the winner is measured in trillions.
For more on the broader AI investment landscape, see our coverage of ElevenLabs' $500M Series D, Waabi's $1B partnership with Uber, and Runway's $315M world models round.
What is Nscale and what does it do?
Nscale is a full-stack AI cloud platform headquartered in London. It owns and operates data centers across Europe and North America, provides GPU compute clusters for training and deploying AI models, and offers software services including serverless inference, fine-tuning, and fleet management. Its customers include OpenAI and Microsoft.
How much has Nscale raised in total?
Including the $2 billion Series C announced March 9, 2026, Nscale has raised approximately $3.6 billion in equity funding (Series A through Series C) plus a $1.4 billion debt facility secured in February 2026 for GPU purchases.
Why is Nvidia investing in Nscale?
Nvidia's investment aligns with its strategic interest in ensuring that its GPUs are deployed in high-efficiency, high-scale infrastructure globally. Nscale has committed to deploying tens of thousands of NVIDIA Blackwell GPUs across its data centers. By taking an equity stake, Nvidia deepens its relationship with a major customer and gains a strategic presence in European AI infrastructure.
What makes Nscale different from AWS or Azure for AI?
Nscale is AI-native, meaning its infrastructure and software stack were designed specifically for AI training and inference workloads rather than adapted from general-purpose cloud computing. It also operates with renewable energy in Norway and Iceland, giving it a cost advantage (CEO claims 10% lower compute costs than competitors) and a sustainability profile that matters to European enterprise and government customers. Its European headquarters and data center footprint also address data sovereignty requirements that US hyperscalers struggle to fully satisfy under EU regulations.
Who are Sheryl Sandberg, Nick Clegg, and Susan Decker, and why do their board appointments matter?
Sheryl Sandberg is the former COO of Meta who built its global advertising platform. Nick Clegg is a former UK Deputy Prime Minister who most recently served as Meta's President of Global Affairs. Susan Decker is the former President of Yahoo. Their appointments signal Nscale's intent to compete at the highest levels of enterprise and government engagement — Sandberg brings enterprise sales expertise, Clegg brings European regulatory and government relationships, and Decker brings experience managing complex, fast-scaling organizations.
Nscale's $2 billion Series C is a landmark moment — not just for the company, but for European AI infrastructure as a category. The $14.6 billion valuation, the Nvidia endorsement, the blue-chip board, and the concrete delivery record across multiple countries position Nscale as the credible European answer to the AI compute buildout that has been dominated by US hyperscalers and, increasingly, by US-based GPU cloud specialists like CoreWeave.
The market context is compelling: AI compute demand is growing faster than infrastructure can be built, energy constraints are creating sustainable advantages for players with access to renewable power, and sovereign AI policy is making European data center capacity a matter of national strategic priority. Nscale is positioned at the intersection of all three dynamics.
The company now has the capital, the partnerships, and — with Sandberg, Clegg, and Decker on the board — the institutional credibility to compete for the largest enterprise and government AI contracts in Europe. Whether it can translate that positioning into enduring hyperscaler-scale revenue is the open question. The next few years of execution will tell that story.
For now, March 9, 2026 marks the day Europe got a serious AI infrastructure champion backed by the world's most important chip company, one of history's most successful technology operators, and the institutions needed to navigate the political economy of AI at scale. That is a story worth paying attention to.
OpenAI's $110 billion funding round — led by Amazon's $50 billion stake — has forced a restructuring of the Microsoft-OpenAI partnership that defined enterprise AI since 2019.
Rowspace exits stealth with $50M from Sequoia and Emergence Capital, targeting private equity firms drowning in deal docs and fragmented data — already live at top PE firms managing over $100 billion in AUM.
German robotics startup Neura Robotics closed approximately €1 billion in funding from Tether Holdings, valuing the company at €4 billion as it prepares to fill nearly €1 billion in existing orders for cognitive humanoid machines.