Saudi Arabia is doubling down on its bid to become a global AI superpower. On October 22, 2026, HUMAIN — the Kingdom's PIF-owned AI company — announced an expanded partnership with NVIDIA to deploy up to 84 megawatts of AI computing capacity, roughly equivalent to 27,000 NVIDIA GPUs.
The announcement, made at NVIDIA's GTC conference in Berlin, marks one of the largest sovereign AI infrastructure commitments in the Middle East this year — and a direct escalation of the Gulf's race to become the region's AI compute hub.
What the deal covers
The planned rollout spans two NVIDIA hardware generations. Phase one covers 14 megawatts of Blackwell systems — about 6,000 GPUs — with deployment starting in the fourth quarter of 2026. Phase two adds a further 70 megawatts of Vera Rubin infrastructure, roughly 21,000 GPUs, planned from mid-2027.
Notably, 50 megawatts of the Vera Rubin capacity is already earmarked for an unnamed US client, subject to final agreements. That detail matters: it suggests HUMAIN is positioning itself not just as a domestic AI champion, but as an exporter of AI compute to American customers.
The 84 MW plan builds on an earlier agreement from May 2025 targeting up to 500 megawatts of AI capacity over five years.
The NEOM connection
Alongside the NVIDIA expansion, HUMAIN and DataVolt signed a preliminary framework with Nebius for up to 50 megawatts of data-center capacity at Oxagon, NEOM's floating industrial city. At LEAP 2026, HUMAIN also announced a strategic partnership with MinIO to build a unified AI platform integrating infrastructure, data, models, and applications, with MinIO serving as its primary data and memory partner.
Why it matters: the Gulf AI race
HUMAIN was launched in May 2025, is owned by Saudi Arabia's Public Investment Fund, and is chaired by Crown Prince Mohammed bin Salman. The company has said it plans to develop about 3 gigawatts of data-center capacity by 2030 and more than 6 gigawatts by 2034.
It is not alone. In the UAE, G42, OpenAI, and Oracle are building Stargate UAE — a 1-gigawatt computing cluster forming the first phase of a planned 5-gigawatt UAE-US AI campus, also with NVIDIA as technology partner.
Both Gulf states face the same constraint: dependence on US export approvals to obtain advanced AI chips, making Washington's technology restrictions an important factor in their expansion plans.
Vera Rubin takes center stage
The deal also underscores how central NVIDIA's next-generation Vera Rubin platform has become. Rubin, the successor to Blackwell, entered full production in early 2026, and NVIDIA claims roughly 5x inference and 3.5x training performance over Blackwell — its own figures, not yet independently verified.
With hyperscalers and now sovereign AI companies lining up for Vera Rubin capacity, the platform is shaping up as the defining AI infrastructure of 2027.
The bottom line
Up to 84 megawatts and 27,000 GPUs is a serious commitment — but in the global AI infrastructure race, it is still a first step toward HUMAIN's 3-gigawatt 2030 target. Whether Saudi Arabia's deep pockets and energy advantage can turn it into a global AI compute supplier will depend on execution, talent, and continued access to cutting-edge chips.
Can the Gulf become the world's next AI superpower — or will power, talent, and chip approvals decide the race? Tell us in the comments.
Comments
Post a Comment