According to media reports citing informed sources, UK-based AI compute power leasing startup Nscale is in discussions with Wall Street institutional investors to raise approximately $20 billion in new funding. This "neo-cloud" startup, focused on AI compute leasing, is pursuing this massive capital injection just over three months after completing its previous funding round. This aggressive fundraising pace, even surpassing that of OpenAI, underscores the persistently surging global demand for AI computing power, compelling Nscale to rapidly "replenish its arsenal" to expand its AI infrastructure. The startup previously closed two highly successful funding rounds in September and October, valued at approximately $11 billion and $4.33 billion respectively. The $11 billion round was confirmed to be the largest Series B funding in European history. Sources indicate that this large-scale neo-cloud service provider specializing in AI compute is collaborating with Wall Street giants Goldman Sachs and JPMorgan on the fundraising efforts. The sources requested anonymity as the discussions involve confidential information.
The British startup Nscale raised over $15 billion combined in its September and October rounds. Shortly before that, the company announced partnerships with "AI chip dominator" NVIDIA and AI application leader OpenAI. The fact that Nscale, having just secured over $15 billion in September-October, is now negotiating a new round of about $20 billion after an interval of "just over three months" reflects the intensely capital-intensive nature of this "neo-cloud/GPU compute leasing and physical delivery" business, requiring continuous capital infusion to secure GPU supply, power, and delivery capacity for entire AI data center core equipment. Sources noted that the financing discussions are ongoing, and the startup may ultimately decide not to proceed with the transaction. Representatives for Nscale, Goldman Sachs, and JPMorgan all declined to comment.
This "neo-cloud" startup, focused on leasing cloud-based AI compute infrastructure, is a typical player in the "AI compute infrastructure layer." Based in the UK, Nscale's overall business model is largely comparable to the cloud AI compute leasing giant CoreWeave, which is often dubbed "NVIDIA's favored son" and boasts a market valuation of $38.5 billion on the US stock market. Some investors are already envisioning this new-generation "AI compute leasing arms dealer" Nscale becoming the "next CoreWeave."
What exactly is Nscale, which is backed by NVIDIA? Nscale was spun off from a cryptocurrency mining enterprise in 2024, joining the so-called "neo-clouds" camp of cloud-based AI compute leasing. It provides enterprises requiring specialized AI computing power or larger-scale AI infrastructure with access to physically delivered or cloud-leased AI compute capacity. Nscale's core revenue logic typically stems from two streams: First, "Long-term/Reserved Compute & Infrastructure Service Contracts," where clients pay for "deterministic capacity" (akin to "booking slots/clusters" for physical delivery or long-term leasing of AI GPU compute clusters), plus fees for services like operations and maintenance, high-performance networking, storage equipment, and cloud hosting delivery. Nscale has publicly disclosed a multi-year physical AI infrastructure service delivery arrangement with Microsoft.
Second, it generates revenue from "On-demand cloud AI compute leasing/Long-term exclusive use" products like Public Cloud, Serverless inference, and Fine-tuning, which are typically billed based on actual usage duration of cloud AI GPU clusters, inference calls, bandwidth/high-performance storage, etc. It is understood that US tech giant Microsoft is investing $10 billion to build a data center in Portugal, advancing the project in collaboration with Nscale and a local Portuguese partner. Microsoft also plans to lease even larger-scale AI compute capacity from Nscale in Norway and the UK.
The "neo-cloud" startup focused on AI compute leasing has made several crucial senior management hires in recent months, which some investors view as very critical for its potential initial public offering (IPO). NScale recruited Lauren Hurwitz from US-based "AI + data center" application leader Palantir Technologies Inc. this month to serve as Chief Operating Officer. In December, the company appointed long-time JPMorgan banker Alice Takhtajan as Chief Financial Officer, and named former long-time Microsoft executive Nidhi Chappell as President of AI Compute Infrastructure at NScale; Chappell previously led Microsoft's Azure AI and high-performance computing infrastructure business.
It is understood that the world's highest-valued company, "AI chip dominator" NVIDIA, has a dual deep relationship with Nscale involving both "investment + business cooperation." Currently public information regarding OpenAI and Nscale mainly indicates "business cooperation," with no clear disclosure of an equity investment by OpenAI in Nscale. The stock prices of Nscale's already-public competitors have been quite volatile in recent months, primarily as investors begin focusing on "AI bubble rhetoric" and questioning the pace of AI infrastructure spending by US tech giants.
CoreWeave Inc. went public on the US stock market in late March 2025 with a fundraising size of $1.57 billion. Although its share price has significantly retreated from its June all-time high of $187 per share, it remains at approximately twice its $40 IPO issue price. Amsterdam-based cloud service provider Nebius Group NV, also publicly traded in the US, has seen its stock price more than triple over the past year, largely as investors acknowledge its high-profile partnerships with tech giants like Microsoft. Nebius Group NV's market capitalization was approximately $24.5 billion at the close of US trading on Thursday.
Nscale, striving to become the next CoreWeave. Nscale CEO Josh Payne stated in a recent declaration: "We are building one of the largest global neo-cloud platforms of its kind, specifically designed to meet the surging demand for AI compute and achieve major breakthroughs at an unprecedented scale." "This will enable Nscale to provide our customer base with scarce and highly sought-after computing capacity, and rapidly accelerate the build-out of secure, compliant, energy-efficient, and high-performance AI compute infrastructure."
The investment and deep collaboration with NVIDIA under Jensen Huang's leadership, coupled with deep cooperation with OpenAI on major European AI infrastructure projects, and collaborations with Microsoft related to compute and data center construction, are all reinforcing its market positioning as an "AI GPU cloud computing infrastructure cluster platform." Based on disclosed information, UK-based Nscale's business model is highly similar to that of "NVIDIA's favored son" CoreWeave. The core of both revolves around providing AI training/inference infrastructure based on either physical AI data center construction/leasing services or a cloud-based leasing system.
Nscale, CoreWeave, and Nebius Group NV can be considered players in the same neocloud (new cloud) arena. CoreWeave has already navigated the path to IPO/public market pricing, while Nscale is in a similar capital market narrative phase of "fundraising - large-scale business expansion - hiring top Wall Street executives in preparation for a US IPO" (at least from external observation and investor expectations). As one of the earliest adopters of cloud leasing for NVIDIA GPUs in the data center space, CoreWeave capitalized early on the wave of demand for AI compute resources in data centers, gaining favor with NVIDIA's venture arm and even securing priority access to the highly sought-after NVIDIA H100/H200 and Blackwell series AI GPUs. This position once forced cloud service giants like Microsoft to lease cloud AI compute resources from CoreWeave, earning it the "NVIDIA's favored son" moniker.
Current global demand for AI compute resources is undoubtedly exhibiting explosive expansion. This is why valuations for leaders in cloud AI compute leasing like Fluidstack and CoreWeave have continued to expand this year. The demand for AI compute resources, closely tied to AI training/inference, has pushed the capacity limits of underlying compute infrastructure clusters. Even with the ongoing construction of large AI data centers, they cannot meet the incredibly strong global demand for computing power. Following Google's major launch of the Gemini 3 AI application ecosystem in late November, this cutting-edge AI software quickly gained global popularity, causing a sudden surge in Google's AI compute demand.
The release of the Gemini 3 series products immediately generated an immense volume of AI token processing, forcing Google to significantly reduce free access limits for Gemini 3 Pro and Nano Banana Pro and impose temporary restrictions even on Pro subscribers. Coupled with recent South Korean trade export data showing持续强劲 demand for HBM memory systems and enterprise-grade SSDs, this further validates Wall Street's assertion that the "AI boom is still in the early construction phase where compute infrastructure supply cannot keep up with demand."
