UK-based AI cloud provider Nscale has filed for a U.S. IPO and secured a landmark \$45 billion compute deal with Anthropic. The moves come against a backdrop of surging demand for power and compute infrastructure driven by increasingly sophisticated AI models.

Key Highlights of the Filing and the $45B Deal with Anthropic

London-based AI infrastructure and cloud provider Nscale filed a Form S-1 registration statement for an initial public offering (IPO) with the U.S. Securities and Exchange Commission (SEC) on September 18, 2026. The company plans to list on the New York Stock Exchange under the ticker symbol “NSCL,” aiming to raise up to \$3 billion. Ahead of its public debut, Nscale clinched a landmark cloud computing agreement with AI powerhouse Anthropic, valued at up to \$45 billion over six years.

Under the agreement, Nscale will reserve approximately 460 megawatts (MW) of power capacity for Anthropic at its Monarch Compute Campus currently under construction in West Virginia. That amount of electricity is equivalent to the power consumed by roughly 345,000 U.S. households, underscoring the colossal scale of infrastructure needed for next-generation AI training. The facility is set to deploy NVIDIA’s cutting-edge NVIDIA Vera Rubin systems, with compute capacity scheduled to begin coming online between late 2027 and early 2028. Please refer to the conceptual diagram below.

Through this long-term contract, Nscale aims to establish a stable revenue foundation and boost its valuation in public markets.

Rapid Growth vs. Capital-Intensive Business Realities

Originally spun off from a crypto-mining operation in early 2024, Nscale rapidly pivoted into a dedicated AI data center operator. According to its Form S-1 filing, revenue for the six months ended June 30, 2026, surged to \$140.6 million—a staggering twelve-fold increase (up 1,252%) from \$10.4 million in the same period a year earlier. The company boasts a total contracted value (TCV) exceeding \$103 billion across active and signed deals, backed by a vast pipeline supporting approximately 461,000 GPUs.

At the same time, heavy upfront investments required for rapid expansion continue to weigh on its finances. Net loss widened to \$1.02 billion in the first half of 2026, compared to \$368.9 million in the prior-year period. This reflects the classic capital-intensive model of the AI data center industry, where massive upfront outlays are required for GPU procurement, land acquisition, power infrastructure, and cooling systems long before contract revenue is fully realized. Cost of revenue reached \$189.6 million, exceeding revenue for the period. Nscale intends to use the capital raised through its IPO to fund these infrastructure buildouts, pay down debt, and support ongoing capital expenditure.

Nscale’s Technological Foundation and Strategic Partnerships

Close Financial Ties with NVIDIA and Off-Grid Infrastructure Tech

A cornerstone of Nscale’s rapid capacity expansion is its deep relationship with semiconductor giant NVIDIA. NVIDIA acquired warrants during Nscale’s 2025 Series B funding round, making it a major shareholder holding more than 5% of non-voting shares. NVIDIA is also slated to provide Nscale with a \$1 billion convertible loan facility. However, Nscale cited its heavy reliance on NVIDIA for GPU supply as a primary risk factor in its IPO filing.

On the technology front, Nscale distinguishes itself through an autonomous infrastructure design that reduces dependence on local power grids. At the Monarch campus in West Virginia, the company employs an on-site microgrid developed jointly with Caterpillar, utilizing natural gas and diesel engines for power generation. This enables the data center to operate without straining local utilities. Furthermore, the campus uses a closed-loop cooling system that eliminates the need for municipal drinking water. Please see the architecture diagram below.

This environmentally conscious, self-sustaining infrastructure design serves as a key differentiator against competitors struggling to secure grid power.

Expanding into Software and Diversifying Partnerships

Nscale is expanding beyond mere land and hardware leasing toward becoming a full-stack AI platform provider. In July 2026, the company agreed to acquire San Francisco–based AI software startup Anyscale—founded by the creators of the open-source distributed computing framework Ray—for \$1.65 billion. By bringing Anyscale’s software layer in-house, Nscale enables customers to manage and orchestrate large-scale GPU workloads efficiently, positioning itself to monetize both hardware and software layers.

Nscale’s customer pipeline extends well beyond hyperscalers and AI labs like Anthropic and Microsoft. In September 2026, the company signed a multiyear deal with humanoid robotics firm Figure. The agreement features an initial compute commitment of \$3.5 billion, with plans to expand beyond \$6 billion, involving up to 100,000 NVIDIA Vera Rubin GPUs. Through these deals, Nscale is successfully diversifying its customer base into robotics and advanced industries, solidifying its long-term growth trajectory.

Challenges in the AI Infrastructure Market and Future Outlook

Customer Concentration Risks and Huge Capital Requirements

Despite its impressive growth trajectory, Nscale faces significant challenges and uncertainties. Chief among them is customer concentration risk. According to the S-1 filing, a single top customer accounted for 52% of total revenue in the first half of 2026. Furthermore, its massive deals—up to \$43.8 billion with Microsoft and up to \$44.6 billion with Anthropic—are subject to strict delivery milestones and service availability requirements, creating risk if implementation hits delays.

Another hurdle is securing the capital required to fulfill these commitments. Executing the \$45 billion Anthropic deal requires enormous capital for GPU hardware and data center construction, yet Nscale noted in its filing that it has not yet secured binding financing commitments for all required equipment and infrastructure. While the company is seeking to raise \$3.5 billion via convertible debt with assistance from Goldman Sachs prior to the IPO, shifts in market conditions could drive up financing costs.

Industry-Wide Impact and Wall Street Focus

Nscale’s IPO filing and massive contract with Anthropic highlight the ongoing, intense global race for AI compute capacity. For Anthropic, which faced capacity bottlenecks and service disruptions due to surging demand for its Claude models earlier this year, a multi-vendor strategy is essential. By spreading workloads across AWS, Google Cloud, and emerging “New Cloud” operators like Nscale, Anthropic mitigates vendor lock-in. With Anthropic eyeing its own public debut in late 2026 or beyond, securing compute infrastructure is vital to supporting its valuation.

On Wall Street, investors are closely watching Nscale’s competitive positioning against rivals like CoreWeave and Nebius, as well as broader AI infrastructure valuations. AI infrastructure providers have become a primary engine of the global IPO market in 2026. Nscale’s post-IPO performance will serve as a bellwether for investor appetite in the sector. Investors will be scrutinizing whether Nscale can sustain its rapid revenue growth while eventually steering its expanding net losses toward profitability.

Reference Pages

• 【SEC Form S-1 Filing】https://www.sec.gov/Archives/edgar/data/2110365/000119312526395475/ck0002110365-20260918.htm
• 【Nscale Monarch Compute Campus】https://www.nscale.com/ai-infrastructure/monarch

【#Nscale #Anthropic #NVIDIA #IPO #AIInfrastructure #DataCenter #CloudComputing】