Skip to content
13 November 2026

Nscale IPO reveals loss surge and uncertain financing

Nscale’s updated S‑1 puts a $1.69 billion loss alongside a 1,344% revenue rise, while its biggest deal still lacks secured funding.

Nscale IPO reveals loss surge and uncertain financing

On November 12, Nscale Limited submitted an amended registration statement for a planned New York Stock Exchange debut. The filing, the first amendment since the original September S-1, replaces earlier projections with a picture of soaring revenue, deeper losses and an unresolved financing puzzle. The company, backed by Nvidia and positioned as a “neocloud” for artificial-intelligence workloads, hopes the market will price its shares under the ticker NSCL, but the numbers raise eyebrows.

The financial snapshot is stark. For the first nine months of 2026, Nscale reported revenue of $307.5 million, a jump of 1,344% compared with the same period a year earlier. That growth stems mainly from signed contracts, not from cash-generating operations. In the same window the firm posted a net loss of $1.69 billion, meaning it spent roughly five dollars for every dollar earned. The loss includes $457 million of fair-value adjustments on convertible notes and warrants, items that disappear once the company becomes public.

Revenue surge versus cash reality

Although the top line looks impressive, the cash flow story tells a different tale. Operating cash flow was positive $1.69 billion in the first half of the year, but $1.6 billion of that figure came from pre-payments: money customers have advanced for capacity that has yet to be built. Calling this a cash-generation success ignores the fact that the funds are tied to future construction and hardware purchases. The filing also lists $24 billion of equipment purchase commitments and an additional $3.5 billion for construction, most of which must be settled within the current and next fiscal years.

Anthropic contract and the financing gap

The centerpiece of Nscale’s order book is a multi-year agreement with AI-model developer Anthropic. The deal, signed on August 25, promises up to $44.6 billion of GPU capacity at the planned Monarch Compute Campus in West Virginia. The campus is slated for an initial 460-megawatt power draw, a scale comparable to a small city. Yet the S-1 explicitly states that Nscale has not secured binding financing for the GPUs and data-center infrastructure needed to honor the contract. In effect, the company holds a massive order but lacks the committed capital to build the required hardware.

To narrow the gap, Nscale has pursued a convertible-note round, reportedly raising $3.36 billion, with Nvidia earmarking $1 billion of that amount. The notes carry a $30 billion valuation cap, signaling the investors’ view of a reasonable price. However, the capital raised falls short of the $31-$32 billion estimated to complete the Monarch build-out, leaving a sizable shortfall that must be covered before the Anthropic timeline begins in late 2027.

Market reaction and outlook

Goldman Sachs, J.P. Morgan and Morgan Stanley are leading the underwriting, but the prospectus offers no price range or fundraising target, leaving the market to guess the ultimate valuation. Some potential backers have voiced concerns that a valuation of $30-$35 billion may be optimistic given the current borrowing environment, higher debt costs for data-center projects, and recent share-price declines at rivals such as CoreWeave. Adding to the uncertainty, Microsoft withdrew from a planned English site and did not renew a Texas contract, reducing the company’s near-term customer base.

Investors now face a classic dilemma: a lightning-fast increase in signed revenue versus a substantial, unfilled financing requirement. The success of the IPO will hinge on whether Nscale can lock down the capital needed for Monarch and convince the market that its pre-payment model can sustain the heavy upfront investment. Until those financing commitments become concrete, the $35 billion valuation remains a projection rather than a market-determined price.

Author

James Carter