Global AI, the U.S. builder of single-tenant, air-gapped data centers, has closed its first-ever debt raise, securing $441 million in a senior secured credit facility led by J.P. Morgan. The financing will expand the company's sovereign AI data center capacity, which it says is in exceptional demand.
The facility, arranged and led by J.P. Morgan, drew participation from a group of lenders the company did not name. The raise comes just days after Nscale, a GPU infrastructure operator, secured a $900 million credit facility from a lender group that also included J.P. Morgan. Two nine-figure debt packages for non-hyperscaler AI capacity landed within days of each other, a sign that lenders are warming to infrastructure outside the major cloud providers.
A first debt raise for a young company
Global AI was founded in 2024 and describes itself as the world's first sovereign AI hyperscaler. The company designs, builds, powers, and operates data centers where one customer occupies the entire environment, physically isolated from other tenants and public networks. Its model spans land, energy, construction, liquid cooling, and GPU-dense compute, targeting nations and enterprises that want AI training and inference inside their own jurisdiction.
Sami Issa, Co-Founder, Director, and CEO of Global AI, said the company is moving faster on capital. "We are significantly accelerating our capital-raising efforts to meet the exceptional demand we are seeing for dedicated, single-tenant, air-gapped AI infrastructure," Issa said.
He pointed to the company's order book as evidence of that demand. "With $6.2 billion in contracted revenues, including $1 billion in contracted revenues already built and delivered to customers, we are scaling rapidly to meet this demand," Issa said.
From Endicott to a gigawatt target
Global AI owns and operates a data center in Endicott, New York, running GB200 NVL72 clusters. The site started with 16 megawatts of initial critical IT load and is ramping to 100 megawatts through 2026. The company plans 250 megawatts of critical capacity for 2027, with a target of 1 gigawatt by the end of 2029.
On March 16, 2026, Global AI completed a GB300 NVL72 deployment, which the company calls the largest such cluster in New York State. It is actively deploying 7,000 GB300 systems across its footprint. The company also plans to roll out NVIDIA's next-generation Vera Rubin NVL72 racks across its U.S. sites, with deployments tied to the 2027 capacity ramp.
Global AI is an NVIDIA Cloud Partner, and NVIDIA provides the hardware for its data centers. NVIDIA also joined a recent Global AI fundraising round, a detail disclosed on December 3, 2025, when the company announced a partnership with HUMAIN, an AI company backed by Saudi Arabia's Public Investment Fund, to develop large-scale AI data centers in the U.S.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
Debt markets open to sovereign AI
The timing of Global AI's raise, so close to Nscale's $900 million facility, suggests a broader trend. Lenders are increasingly financing AI infrastructure that sits outside the hyperscaler giants like AWS, Google, and Microsoft. Both deals involved J.P. Morgan, which appears to be taking a leading role in this emerging corner of the market.
For Global AI, the $441 million facility marks a shift from equity to debt. The company had previously relied on fundraising rounds, including one that drew NVIDIA. Now it has a senior secured credit facility, a structure that gives lenders priority claims on assets, which may make it easier to raise larger sums later.
The company's contracted revenues, at $6.2 billion, far exceed the size of its first debt raise. That gap between contracted demand and available capital helps explain why Issa says the company is accelerating its capital-raising efforts. The $1 billion in contracted revenues already built and delivered shows the company is converting orders into operating assets.
What comes next
Global AI's expansion plans hinge on its ability to keep building. The Endicott facility is the anchor, but the 1 gigawatt target by the end of 2029 will require multiple sites. The partnership with HUMAIN is expected to add large-scale U.S. data centers, though the companies have not disclosed specific locations or timelines.
The Vera Rubin NVL72 rollout, planned for 2027, will bring NVIDIA's next-generation rack systems into Global AI's facilities. The company is also working through its 7,000 GB300 systems, which are already in active deployment. Those systems, along with the GB200 clusters in Endicott, form the compute base for its sovereign AI customers.
The debt market's appetite for this kind of infrastructure appears to be growing. Nscale's $900 million facility and Global AI's $441 million raise, both within the same week, point to a financing pipeline for non-hyperscaler AI capacity that did not exist a year ago. Whether that pipeline stays open will depend on how quickly these companies can turn contracted revenue into delivered data centers.