World AI, a U.S. firm that builds single-tenant, air-gapped knowledge facilities for governments and enterprises, has closed $441 million in financing, its first-ever debt increase. The senior secured credit score facility was led and organized by J.P. Morgan, with participation from a bunch of lenders the corporate didn’t title.
The announcement lands the identical week one other GPU infrastructure operator, Nscale, lined up a $900 million credit score facility from a lender group that additionally included J.P. Morgan — two nine-figure debt packages for non-hyperscaler AI capability inside days of one another.
Co-Founder, Director, and CEO Sami Issa tied the increase to contracted demand. “We’re considerably accelerating our capital-raising efforts to satisfy the distinctive demand we’re seeing for devoted, single-tenant, air-gapped AI infrastructure,” Issa stated. “With $6.2 billion in contracted revenues, together with $1 billion in contracted revenues already constructed and delivered to prospects, we’re scaling quickly to satisfy this demand.”
What World AI is definitely constructing
Based in 2024, the corporate describes itself because the world’s first sovereign AI hyperscaler: it designs, builds, powers, and operates knowledge facilities the place one buyer occupies the whole surroundings, bodily remoted from different tenants and from public networks. Its mannequin spans land, vitality, development, liquid cooling, and GPU-dense compute, geared toward nations and enterprises that need AI coaching and inference inside their very own jurisdiction reasonably than on shared public cloud.
The {hardware} program is NVIDIA-based. The corporate, an NVIDIA Cloud Associate, owns and operates a knowledge heart in Endicott, New York, operating GB200 NVL72 clusters. On December 3, 2025, it introduced a partnership with HUMAIN, the AI firm backed by Saudi Arabia’s Public Funding Fund, to develop large-scale AI knowledge facilities in the US; that launch additionally disclosed that NVIDIA had joined a current World AI fundraising spherical.
On March 16, 2026, the corporate stated it had accomplished a GB300 NVL72 deployment it calls the biggest such cluster in New York State, is actively deploying 7,000 GB300 methods, and plans to roll out NVIDIA’s next-generation Vera Rubin NVL72 racks throughout its U.S. footprint.
The capability roadmap behind the increase
The debt comes in opposition to a buildout the corporate has already scheduled publicly: 16 megawatts of preliminary essential IT load at Endicott, ramping to 100 megawatts by means of 2026, with 250 megawatts deliberate for 2027 and 1 gigawatt of essential capability focused by the top of 2029.
For an organization based in 2024, the $441 million senior secured facility in opposition to $6.2 billion in contracted income is its first-ever debt increase, following earlier fundraising rounds that included NVIDIA. The acknowledged subsequent milestone is the ramp to 100 megawatts at Endicott by the top of 2026, adopted by the Vera Rubin NVL72 deployments and the 250-megawatt mark in 2027.
