Startups

Nscale targets $3B IPO as AI data center demand surges

The London-based infrastructure builder holds $51 billion in contracted revenue and plans to expand capacity from 831 megawatts to 11 gigawatts.

Omega Editorial· August 22, 2026· 3 min read

Nscale Global Holdings Ltd. is preparing to raise $3 billion through an initial public offering that could launch as soon as next month, according to a report from Bloomberg citing sources familiar with the matter.

The London-based AI data center builder has tapped Goldman Sachs Group Inc. and JPMorgan Chase & Co. to manage the listing on a U.S. stock exchange. The company most recently carried a $14.6 billion valuation following a March funding round that included investments from Nvidia Corp. and Nokia Corp.

Why it matters

Nscale's IPO timing reflects the intense capital requirements of AI infrastructure buildout and growing investor appetite for companies positioned between hyperscalers and chip manufacturers. With $51 billion in contracted revenue already secured, the company represents a test case for whether specialized data center operators can command public market valuations that reflect long-term capacity commitments rather than current operations.

Aggressive expansion plans

Nscale opened its first data center in Norway last year and now operates AI infrastructure projects in more than a dozen locations globally. The company aims to scale its capacity from 831 megawatts today to approximately 11 gigawatts, according to Bloomberg's sources.

The centerpiece of this expansion is a 2,250-acre campus in West Virginia that Nscale estimates could accommodate more than eight gigawatts of computing capacity. The site will feature a dedicated electrical grid with on-site power generation infrastructure.

Microsoft Corp. serves as the anchor tenant for the West Virginia facility, having commissioned 1.35 gigawatts of capacity in March. Nscale will deploy Nvidia's Vera Rubin NVL72 systems for that contract, with each appliance containing 72 of the chipmaker's latest Rubin graphics processing units.

In a separate agreement reportedly valued at $14 billion, Nscale will host 300,000 previous-generation Blackwell Ultra chips for Microsoft across four additional locations.

Software layer adds differentiation

Beyond raw computing capacity, Nscale offers cloud services designed to simplify hardware utilization. The company provides managed versions of Kubernetes and Slurm, frameworks that help optimize AI workload infrastructure usage. It also offers a prompt engineering tool aimed at improving model response quality.

Last month, Nscale acquired startup Anyscale Inc. for $1.65 billion, gaining a commercial version of Ray, an open-source tool for optimizing AI clusters. The Anyscale product adds ease-of-use features and observability capabilities to the base Ray framework.

Competitive landscape

Nscale is one of two AI data center operators moving toward public markets in the near term. Earlier this month, rival Switch Inc. filed for an IPO that could reportedly value the company at $50 billion.

Bloomberg first reported the details of Nscale's IPO preparations.

#nscale#ipo#ai data centers#microsoft#nvidia#infrastructure

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Startups

Startups· 2 min read

Anthropic IPO Filing to Flag AI Backlash as Key Risk Factor

The Claude maker is preparing to go public as public opposition to AI and data centers intensifies ahead of midterm elections.

Via AI Watch · Aug 21, 2026
Startups· 3 min read

Micro1 Hits $500M Run Rate as AI Labs Race for Training Data

The data-labeling startup quintupled revenue in eight months by connecting domain experts with AI companies hungry for high-quality datasets.

Via AI Watch · Aug 21, 2026
Startups· 2 min read

AI Investment Cycle Mirrors Check Kiting Fraud Structure

Circular funding patterns among AI companies create timing risks reminiscent of an old banking scam, raising concerns about market stability.

Via AI Watch · Aug 20, 2026