Nscale has grown from a young British AI infrastructure company into a business carrying more than $100 billion in contracted value. Now, Wall Street must decide how much of that future demand deserves to be reflected in its public-market valuation.
The Nscale IPO moved forward after the London-based company filed for a New York Stock Exchange listing under the ticker NSCL. The number of shares and proposed price range have not yet been disclosed.
The filing arrives less than three years after Nscale began building its vertically integrated AI infrastructure business. It combines computing capacity, data centers, power resources, and software for companies training and running artificial intelligence models.
Its growth has been extraordinary, but so has the amount of money required to support that expansion. Revenue reached $140.6 million during the first six months of 2026, compared with $10.4 million one year earlier.
That represents revenue growth of more than 1,250% within twelve months. However, Nscale also reported a $1.02 billion net loss, rising from $368.9 million during the comparable 2025 period.
Those figures capture the central question surrounding the Nscale IPO. Investors must decide whether enormous contracted demand can justify today’s spending before much of that future revenue actually reaches the income statement.
Nscale IPO Puts a $103 Billion Backlog Under Scrutiny
Nscale has expanded rapidly by securing the expensive ingredients required to operate modern AI infrastructure. Those include Nvidia GPUs, large data centers, electricity supplies, networking equipment, and the capital required to bring everything online.
The company says it operates across 14 regions and has developed a power pipeline exceeding 10 gigawatts. Its total contracted value has grown beyond $103 billion after standing at roughly $100 million only two-and-a-half years earlier.
That contracted business is likely to become one of the strongest arguments supporting the Nscale IPO. AI companies increasingly need enormous computing capacity, while securing enough GPUs and electricity remains difficult in many markets.
Anthropic represents one of the clearest examples of that demand. The AI company recently agreed to spend $45 billion renting computing capacity from Nscale’s planned West Virginia campus.
Microsoft and Anthropic are also expected to become major customers during future periods, according to Nscale’s prospectus. These relationships could provide the long-term utilization needed to support its rapidly expanding infrastructure footprint.
Still, contracted value and recognized revenue are fundamentally different numbers. Infrastructure must be built, GPUs installed, power secured, and customers served before much of that contracted value becomes actual revenue.
Customer concentration adds another issue for investors considering the Nscale IPO. The company’s largest customer generated 52% of its first-half revenue, leaving current performance heavily dependent on a relatively small customer base.
That concentration could decline as more contracted customers begin using Nscale infrastructure. However, it also demonstrates how early the company remains relative to the enormous scale suggested by its backlog.
Nscale’s financing activity further illustrates the capital required to deliver those commitments. The company recently agreed to issue $3.1 billion in convertible bonds, including $1 billion purchased by Nvidia.
The company had already raised $2 billion in a Series C round completed in March. That financing valued Nscale at $14.6 billion and included backing from Nvidia, Dell, Nokia, Lenovo, Citadel, and other investors.
Nscale has also expanded beyond simply renting access to GPUs inside third-party facilities. Its strategy increasingly involves controlling the data centers, power infrastructure, computing hardware, and software connecting those resources.
In March, the company agreed to acquire American Intelligence & Power Corporation and its Monarch Compute Campus in West Virginia. The site has a potential power runway exceeding eight gigawatts, according to Nscale.
That infrastructure-first approach could give Nscale greater control over scarce power and computing capacity. However, owning more of the stack also requires significantly more capital before customer revenue catches up.
Wall Street Must Decide How Much Future AI Demand Is Worth
The Nscale IPO arrives during a complicated period for companies supporting the artificial intelligence boom. Demand for computing capacity remains strong, but public investors have become more selective about the economics behind AI infrastructure.
Nscale competes with companies including CoreWeave, Nebius, Crusoe, and Lambda for customers needing powerful GPU clusters. All of them face similar challenges involving equipment costs, electricity availability, financing, construction, and rapidly changing technology.
Nscale’s advantage could come from securing customers before completing much of its infrastructure buildout. Founder and CEO Josh Payne has described the company’s strategy as building infrastructure against contracted customer demand rather than constructing capacity speculatively.
That model gives Nscale greater visibility into future utilization before committing billions to new facilities. Yet execution becomes critical because those contracts must eventually translate into operational data centers and recurring revenue.
Public investors will therefore need to look beyond headline growth when assessing the Nscale IPO. They will be pricing Nscale’s ability to turn a massive pipeline of contracts into revenue while controlling losses and capital requirements.
The $1.02 billion first-half loss shows how expensive that conversion process has already become. Nscale is spending heavily today on infrastructure that management expects customers to use over several years.
That financial structure is common across the emerging neocloud industry, where businesses invest enormous amounts upfront before recognizing revenue gradually. The challenge comes when construction delays, financing costs, customer concentration, or technological changes disrupt those assumptions.
Nscale also enters public markets after its private valuation climbed rapidly. The company reached a $14.6 billion valuation during its March Series C, only months after earlier funding rounds accelerated its global expansion.
Reuters reported that the company is considering a multibillion-dollar listing and could seek a valuation around $30 billion, citing CNBC. Nscale itself has not disclosed an IPO price range or confirmed a final target valuation.
Goldman Sachs, JPMorgan, and Morgan Stanley are leading the proposed offering, alongside a large group of additional banks. Nscale has applied to trade on the NYSE using the ticker NSCL.
The Nscale IPO could therefore become an important test for the wider AI infrastructure sector. Strong demand would suggest investors remain willing to finance companies carrying huge upfront costs when long-term compute contracts provide enough visibility.
Greater caution would send a different message about how markets value contracted AI demand compared with revenue already appearing today. That distinction matters as billions continue flowing into data centers, GPUs, power generation, and specialized cloud providers.
Nscale has already demonstrated that customers are prepared to make enormous long-term commitments for computing capacity. Its next challenge is convincing public investors that those commitments can become durable revenue faster than its infrastructure spending creates financial pressure.
That makes the Nscale IPO about more than one fast-growing British company reaching Wall Street. It will help show how much investors are prepared to pay today for an AI infrastructure market whose biggest revenues still sit years into the future.
