Nscale Seeks $3.5 Billion in Pre-IPO Capital for AI Compute
Cloud provider targets convertible debt and Nvidia backing ahead of public debut.
British AI infrastructure provider Nscale is seeking $3.5 billion in fresh pre-IPO financing to expand its hyperscale compute cluster capacity. The funding round includes $1.5 billion in convertible notes and $2 billion in direct backing from chipmaker Nvidia as compute demand surges.
Key details
Nscale, founded in 2024, is preparing for an initial public offering as early as September 2026. Ahead of its public listing, the company is negotiating $1.5 billion in convertible debt alongside an additional $2 billion investment from Nvidia. Nvidia previously participated in Nscale's $1.1 billion Series B round in March 2026, which followed a $155 million Series A in December 2024.
The capital expansion is anchored by massive commercial commitments, including a recently signed six-year, $45 billion compute agreement with Anthropic. Nscale has reportedly projected over $103 billion in contracted future revenue based on signed customer leases across its global data center portfolio, which includes major compute hubs in West Virginia and Texas.
Why this matters
The multi-billion-dollar financing round highlights the escalating capital requirements for specialist AI cloud providers competing against incumbent hyperscalers. Securing $3.5 billion in pre-IPO liquidity enables Nscale to pre-purchase advanced accelerator hardware and secure long-term electrical power capacity at gigawatt scale.
Context
Specialized AI infrastructure providers have emerged as critical intermediaries between chip manufacturers and model developers. With major frontier labs requiring tens of thousands of GPUs for training and inference, specialized cloud operators rely on heavy debt financing and strategic hardware partnerships to finance massive upfront infrastructure outlays.
What happens next
Nscale's proposed financing structure and initial public offering are expected to move forward in late September 2026, subject to final investor terms. The company will deploy the proceeds toward hardware acquisitions, including up to 100,000 Nvidia GPUs, and expanding its power and facility footprint.
Source: TechCrunch Published on AI Usage Global, author: AUG Bot
