Humain and Nvidia broaden AI partnership with 84MW computing expansion

Saudi Arabia’s Humain has broadened its collaboration with Nvidia, the US chipmaker, to build up to 84MW of AI computing capacity. The plan includes infrastructure for an unidentified US client, as the kingdom intensifies its efforts to establish itself as a global AI centre.

Expanded AI computing partnership

The deal was unveiled at Nvidia’s GTC conference in Berlin on Thursday. It involves approximately 27,000 advanced GPUs. Deployments are set to begin later in 2026, with additional capacity scheduled for 2027.

This expansion follows an earlier partnership announced in May 2025 during a visit by then-US President Donald Trump to Riyadh. At that time, Nvidia said it would supply Humain with hundreds of thousands of advanced AI chips over five years. The first phase included 18,000 Blackwell processors, as part of a wider plan to develop up to 500MW of AI computing capacity within five years.

Deployment timeline and infrastructure details

The initial stage of the new agreement will consist of 14MW of Nvidia’s Blackwell HGX B300 systems, roughly 6,000 GPUs. Deployment is expected to begin in the fourth quarter of 2026. This infrastructure will serve Humain Cloud clients, which include AI firms, government bodies, and businesses.

An additional 70MW of Nvidia’s next-generation Vera Rubin infrastructure, representing about 21,000 GPUs, is planned from mid-2027. Of that capacity, 50MW will be allocated to advanced AI workloads for an unnamed US customer, with deployment targeted for June 2027. The remaining 20MW will support Humain Cloud, with deployment expected to start in July or August 2027.

The companies noted that the Vera Rubin projects are still subject to final agreements and deployment schedules.

GPUs are powerful processors used for training and operating AI systems. Blackwell is Nvidia’s current generation of AI chips, while Vera Rubin is its successor, engineered to manage more complex workloads with greater speed and efficiency.

Strategic context and competition

“We are moving quickly to bring the next generations of Nvidia-accelerated computing into Saudi Arabia at meaningful scale,” said Tareq Amin, chief executive of Humain.

Humain was launched in May 2025 and is chaired by Saudi Crown Prince Mohammed bin Salman. It is owned by the kingdom’s Public Investment Fund and aims to develop about 3GW of data centre capacity by 2030 and over 6GW by 2034.

Saudi Arabia is vying with the UAE to draw AI investment. In Abu Dhabi, G42, OpenAI, and Oracle are developing Stargate UAE, a 1GW computing cluster that represents the first phase of a planned 5GW UAE-US AI campus, with Nvidia also serving as a technology partner.

Both Gulf nations rely on US export approvals to acquire advanced AI chips, making Washington’s technology restrictions a key factor in their expansion strategies.

Conference postponement and security incident

The latest announcement comes days after Saudi Arabia postponed its flagship Future Investment Initiative conference, originally scheduled for 26-29 October, to next year. The postponement followed a Houthi attack on Riyadh’s King Khalid International Airport on 10 October that killed at least 13 people and injured over 300.

Amin was among the speakers scheduled to attend the gathering, which has become a major platform for announcing investment agreements in Saudi Arabia’s technology and infrastructure sectors.

The FII Institute says the annual conference has helped secure $250 billion in signed agreements and investment pledges since its inception.