Nscale Closes $3B Debt Financing: Texas 185MW and North Carolina 40MW AI Supercomputing Campuses Under Full Construction

Introduction

In September 2026, AI infrastructure hyperscaler Nscale closed approximately $3 billion in senior secured delayed-draw term loan financing to build two major AI data center campuses in Texas and North Carolina. This follows Nscale's $2 billion Series C (valuing the company at $14.6 billion) in March 2026 and a $900 million revolving credit facility in July 2026 — the company's third major financing in just six months, bringing cumulative capital raised to over $5.9 billion.

Financing Structure and Scale

$3 Billion Debt Financing Details

This financing consists of two senior secured delayed-draw term loan facilities, with J.P. Morgan and Goldman Sachs serving as joint lead arrangers and bookrunners, and carrying investment-grade ratings:

Texas Ward County Facility: Up to $1.85 billion for a purpose-built, high-density AI campus designed for approximately 275 megawatts (MW) of IT load

North Carolina Madison County Facility: Up to $1.2 billion for a 96-acre colocation site with capacity of up to 40 megawatts (MW)

2026 Financing Timeline

Two AI Supercomputing Campus Details

Texas Ward County: 275MW High-Density AI Campus

The Texas facility is Nscale's most ambitious project, with a designed IT load of 275 megawatts and plans to deploy NVIDIA's latest generation hardware:

Hardware Configuration:

• NVIDIA Blackwell Ultra GPU systems
• NVIDIA Vera Rubin systems (next-generation integrated GPU+CPU architecture)

Cooling Technology:

• Closed-loop direct liquid cooling
• Rear-door heat exchangers

The heat generated by high-density AI computing clusters far exceeds that of traditional data centers, making the liquid cooling technology at the Texas facility critical to managing this challenge. A 275MW IT load is equivalent to the electricity consumption of approximately 200,000 average homes, illustrating the staggering scale of AI infrastructure construction.

North Carolina Madison County: 40MW Colocation Facility

The North Carolina facility takes a different approach, focusing on retrofitting existing data center capacity to meet high-performance AI workload demands:

• Area: 96 acres
• Capacity: Up to 40MW IT load
• Strategy: Retrofitting existing facilities rather than building from scratch

This "retrofit" strategy offers advantages in cost and time efficiency, enabling faster upgrades of existing infrastructure to AI-ready status.

Nscale's Market Positioning

AI Infrastructure Hyperscaler

Nscale positions itself as an infrastructure provider for the AI era, with a core thesis that infrastructure availability is currently the primary constraint on the scaling of global AI systems.

Against the backdrop of explosive growth in AI model training and inference demand, data centers capable of providing large-scale, high-density GPU computing power have become scarce resources. By rapidly building high-density AI data centers, Nscale aims to occupy a favorable position in this supply-demand gap.

IPO Plans

According to August 2026 reports, Nscale has initiated plans for an IPO aiming to raise as much as $3 billion, with a potential timeline as early as September 2026. If successful, this would be a significant IPO event in the AI infrastructure sector, further validating market demand for AI computing infrastructure.

Energy Challenges and Solutions

Power Shortage Issues

Morgan Stanley projects that the U.S. power shortfall could reach as high as 18 gigawatts (GW) through 2028. The rapid expansion of AI data centers is one of the key drivers of this shortage.

Nscale's choice of Ward County, Texas for its facility is partly due to the relatively flexible power market structure of the Texas grid (ERCOT) and Texas's abundant renewable energy resources (wind and solar).

The Importance of Liquid Cooling

High-density AI computing clusters (particularly NVIDIA Blackwell Ultra and Vera Rubin systems) have power density far exceeding traditional servers, making traditional air cooling insufficient. The closed-loop direct liquid cooling technology Nscale employs can remove heat more efficiently while reducing overall energy consumption.

Impact on the Asia-Pacific Region

AI Computing Resource Competition: Nscale's large-scale construction in the U.S. intensifies global competition for AI computing resources. Asia-Pacific AI companies and research institutions need to pay close attention to the accessibility and cost of computing resources.

Local Infrastructure Demand: Nscale's successful model provides a reference for AI infrastructure investment in Asia-Pacific. Japan, Singapore, Australia, and other locations already have similar large-scale AI data center investment plans underway.

Supply Chain Impact: Nscale's large-scale procurement of NVIDIA Blackwell Ultra and Vera Rubin systems may affect global GPU supply chains, in turn impacting GPU procurement plans for Asia-Pacific enterprises.

Industry Trend: The AI Infrastructure Investment Wave

Nscale's financing is a microcosm of the 2026 AI infrastructure investment wave. According to Crunchbase data, global venture funding reached a record $510 billion in H1 2026, with AI-related companies capturing over 70% of global startup capital.

Key investors in the AI infrastructure sector include:

• Sovereign Wealth Funds: Qatar Investment Authority, Abu Dhabi Investment Authority, etc.
• Traditional Tech Giants: NVIDIA, Dell, Lenovo as strategic investors
• Traditional Financial Institutions: J.P. Morgan, Goldman Sachs providing debt financing

Conclusion

Nscale's $3 billion debt financing is an important milestone in the 2026 AI infrastructure investment wave. The construction of the Texas 275MW and North Carolina 40MW campuses will provide critical computing infrastructure support for the U.S. AI industry. In the context of an accelerating global AI race, those who can secure sufficient computing resources will hold a strategic advantage in the AI era.