Saudi artificial intelligence company Humain has expanded its partnership with Nvidia to develop up to 84 megawatts of AI computing capacity in the kingdom, using about 27,000 of the US chipmaker’s graphics processing units.
The companies announced the plan at Nvidia’s GTC conference in Berlin on Thursday, The National reported.
The first phase covers 14MW of Nvidia Blackwell HGX B300 systems, or about 6,000 GPUs. Deployment is due to begin in the fourth quarter of 2026. The capacity will serve customers of Humain Cloud, including AI companies, government entities and local organizations.
A second phase calls for 70MW of Nvidia’s next-generation Vera Rubin infrastructure, or about 21,000 GPUs, from mid-2027.
Of that, 50MW will run frontier-scale AI workloads for a US-based client that the companies did not name. Deployment of that block is targeted to begin in June 2027.
The remaining 20MW will support Humain Cloud, with deployment expected to start in July or August 2027. The Vera Rubin phases remain subject to final agreements and deployment plans.
“We are moving quickly to bring the next generations of Nvidia accelerated computing into Saudi Arabia at meaningful scale,” Tareq Amin, chief executive of Humain, said.
Humain is owned by the Public Investment Fund, Saudi Arabia’s sovereign wealth fund. It plans to build about 3 gigawatts of data center capacity by 2030 and more than 6GW by 2034.
The deal builds on a partnership announced in May 2025, when Nvidia unveiled plans to supply Humain with advanced AI chips over five years. The first phase of that plan involved 18,000 Blackwell processors.
