Nscale Secures $3B for Nvidia Blackwell and Rubin GPUs

Nscale has closed approximately $3 billion in aggregate commitments across two senior secured delayed draw term loan facilities. Announced on August 31, 2026, the financing will fund two major AI infrastructure projects in the United States. The first facility provides up to $1.85 billion for Nscale's Ward County, Texas campus, while the second provides up to $1.2 billion for its Madison, North Carolina site. Both facilities received investment-grade ratings with stable outlooks.

The $1.85 billion Texas loan, issued through Nscale Ward County Borrower SPV, LLC, will fund the deployment of Nvidia GB300 (Blackwell Ultra) and VR200 (Vera Rubin) systems. This purpose-built site is designed to support roughly 275 megawatts of IT load, utilizing closed-loop direct liquid cooling paired with rear-door heat exchangers to handle high-density compute workloads. In contrast, the $1.2 billion North Carolina loan will fund a retrofit of an existing 96-acre colocation property in Madison. This site has up to 40 megawatts of capacity, and the capital will cover physical upgrades, GPU infrastructure, and networking.

Both facilities are structured as delayed draw term loans, allowing Nscale to access capital incrementally as deployment milestones are met. J.P. Morgan and Goldman Sachs served as joint lead arrangers, joint bookrunners, and co-structuring agents. J.P. Morgan led the Texas facility, while Goldman Sachs led the North Carolina project. This deal follows a busy year of financing for Nscale, which closed a $900 million revolving credit facility in July 2026 and a $1.4 billion GPU-backed loan in February 2026, building on a $1.1 billion Series B and a $433 million Pre-Series C SAFE.

For AI practitioners and enterprise developers, Nscale's massive infrastructure expansion translates directly into broader, more reliable access to cutting-edge hardware. By securing the capital to deploy high-density Nvidia Blackwell Ultra and Vera Rubin systems at scale, the company is positioning itself to offer the massive compute power required for training and serving next-generation foundation models. The combination of advanced liquid cooling in Texas and retrofitted colocation space in North Carolina ensures that developers can access diverse, high-performance hosting environments tailored to intensive AI workloads.