The news
Nscale, a London-based company that builds data centers and rents AI computing capacity, filed a registration statement with the SEC on September 18 for a U.S. initial public offering. The Nscale IPO would list the company on the New York Stock Exchange under the ticker NSCL. The share count and price range have not been set.
The filing shows a business growing quickly from a small base. Revenue for the six months ended June 30, 2026, was $140.6 million, compared with $10.4 million a year earlier. Full-year 2025 revenue was $33.0 million. Losses are growing faster in dollar terms: the net loss for the first half of 2026 was $1.02 billion, up from $368.9 million, and adjusted EBITDA was negative $199.2 million. Investing.com’s review of the filing attributes a large share of the loss to $457.1 million in fair value adjustments.
The company’s pitch rests on long-term, take-or-pay contracts. Nscale said it had $2.6 billion in active total contract value and $103.4 billion in active and contracted value as of August 31, with a weighted average contract life of about 5.7 years. It reported about 25,000 active GPUs and 461,000 active and contracted GPUs, across 5 active and 12 contracted sites.
The prospectus lists a September 15 subscription agreement for at least $3.1 billion in convertible notes, including $1.0 billion to be issued to Nvidia, convertible into shares at the IPO. On September 25, Nscale announced $3.36 billion in pre-IPO financing led by Third Point, including Nvidia’s $1 billion. Earlier private rounds raised more than $3.3 billion. The filing warns a substantial portion of revenue comes from a limited number of customers. Goldman Sachs, J.P. Morgan and Morgan Stanley are lead bookrunners.
The numbers
- H1 2026 revenue
- $140.6M (vs. $10.4M)
- H1 2026 net loss
- $1.02B
- Active + contracted contract value
- $103.4B (as of Aug. 31)
- Nvidia investment commitment
- $1.0B (of $3.36B pre-IPO round)
Why CEOs should care
Nscale is one of the “neocloud” providers that sit between chipmakers and AI developers. Its S-1 is a rare public look at their economics: most of the value is in future contracts, while today’s revenue is small relative to losses and capital needs. Public investors will now put a price on that model, which could influence how lenders and private investors fund the next wave of data center builds.
For companies buying AI compute, the filing is a reminder to assess counterparties carefully. A provider with concentrated customers, heavy losses and long build schedules carries execution risk. Multi-year capacity commitments should come with clear delivery milestones, remedies and exit options.
The bigger picture
Nvidia’s $1 billion commitment means the chipmaker is helping finance a company whose clusters run on its Grace Blackwell and Vera Rubin GPUs, per the prospectus.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error
Companies in this story








