Nvidia CEO Jensen Huang’s earnings call namechecked Saudi AI company Humain three times. Here’s why.

Humain is barely six months old, but it is rapidly becoming a major force in the global build out of AI infrastructure. Founded by Saudi Arabia’s Crown Prince Mohammed bin Salman and backed by the nation’s $1 trillion sovereign wealth fund, the Public Investment Fund, Humain has ambitions to supply 6% of the world’s AI computing power by 2034, which would make it the world’s third largest AI data center provider behind the U.S. and China.

Huang’s mentions of Humain on Nvidia’s earnings call come a day after the CEO attended a state dinner at the White House for the Crown Prince, who is visiting the U.S. for the first time since 2018. Coinciding with the visit, Humain announced a deal with Nvidia and Amazon to put 150,000 of Nvidia’s chips, including some of its state-of-the-art Grace Blackwell 300s, in data centers in a new “AI Zone” being built in the Saudi capital Riyadh.

The company also signed a landmark deal with xAI to build a 500 megawatt data center for the company in Saudi Arabia. Nvidia will supply the chips for that data center too.

“Because of our deep partnership with Elon and xAI, we were able to bring that opportunity to Saudi Arabia, to the KSA, so that Humain could also be hosting opportunity for xAI,” Huang said on the earnings call.

Under the leadership of former Aramco executive Tareq Amin, Humain is aiming to be a “full stack” AI company, controlling not just the data centers on which AI models are run, but also building models itself. It has trained and launched a large language model, called ALAM, that was designed to perform better than competitors at Arabic language tasks, as well as avoiding culturally and politically sensitive topics. It has also launched an AI-native laptop and an AI operating system called Humain One.

But Humain’s biggest impact may be as an AI infrastructure builder, creating data centers that it leases to other cloud hyperscalers or AI companies. Saudi Arabia believes its energy resources—including abundant solar power as well as oil and gas—as well as the ease of permitting and construction in the kingdom, mean that it will be able to serve AI software for 30% less than what similar processing would cost in the U.S. The country also has robust fiber optic connections to other countries.

That could make Humain the preferred AI provider for much of the Middle East and Asia, as well as possibly drawing workloads from even further afield.

Who else in the Middle East wants to be an AI hub?

Saudi Arabia is not alone in trying to establish itself as a “third pole” of AI development outside of the U.S. and China. Its regional rival the United Arab Emirates has similar ambitions. Through its own sovereign wealth funds, the UAE has backed G42, a company that is also pursuing a “full stack” approach to AI development.

G42 has been around since 2018 and had a head start on Humain in creating large data centers for generative AI models. But U.S. national security officials under the Biden Administration had raised concerns about G42’s connections to Chinese companies, and had held up exports of Nvidia’s advanced AI chips to the company. These officials worried that the AI technology might leak to Chinese firms. A $1.5 billion investment from Microsoft in April 2024, brokered in part by U.S. government, was supposed to clear the way for G42 to receive Nvidia chips, but both companies complained that the U.S. Commerce Department was slow to approve exports of Nvidia chips to G42 even after the deal.

Some national security experts have raised similar concerns about Humain, since Saudi Arabia, while a U.S. ally, also has defense technology transfer agreements with China. And some Saudi companies, including oil giant Aramco, have been vocal about their use of AI models developed by Chinese companies, such as DeepSeek.

But the Commerce Department just this week approved the export of tens of thousands of Nvidia GPUs to both Saudi Arabia and the UAE.

Meanwhile, Humain has been signing deals with other AI chip providers besides Nvidia. It struck a $10 billion deal with Nvidia’s chief rival AMD to deploy 500 megawatts of AI compute based on AMD’s chips within the next five years. It signed a partnership with Qualcomm to use its AI200 and AI250 AI chips for 200 megawatts of computing capacity, starting in 2026. It has also partnered with AI chip startup Groq.