Riyadh, UAE join global AI race in full force, pile Nvidia superchips

By Al Mayadeen English
Source: Financial Times
15 Aug 2023 18:47

The two Gulf states are currently in the middle of major projects that will turn the countries into AI powerhouses.

An illustration of an Nvidia processor (Nvidia image release)

Saudi Arabia and the UAE have joined the globally heated race to secure high-performance Nvidia chips, a critical component for developing artificial intelligence AI software, as the rich Gulf states aim to pioneer the exponentially growing field, the Financial Times reported on Tuesday.

Riyadh bought at least 3,000 of Nvidia’s H100 chips - a $40,000 processor described by the tech firm's chief Jensen Huang as “the world’s first computer [chip] designed for generative AI,” FT said according to unnamed sources. The order is expected to arrive by 2023 end and is estimated to cost the kingdom around $120 million.

The purchase was made through the state-owned research institution King Abdullah University of Science and Technology (Kaust).

To understand the significance of this move, estimates suggest that OpenAI trained its advanced GPT-3 model on 1,024 A100 chips, the predecessor of the chip currently ordered by Riyadh. The whole model was trained in just one month.

On the other hand, the UAE also managed to put its hands on thousands of Nvidia chips and has already developed its own open-source language model - Falcon - at the public Technology Innovation Institute in Masdar City, Abu Dhabi.

“The UAE has made a decision that it wants to . . . own and control its own computational power and talent, have their own platforms, and not be dependent on the Chinese or the Americans,” said a source familiar with Abu Dhabi’s policy.

“Importantly, they have the capital to do it, and they have the energy resources to do that and are attracting the best global talent as well,” they added.

According to FT newspaper, Kaust possesses at least 200 A100 Nvidia chips and is building a supercomputer dubbed Shaheen III, which is expected to launch later this year.

The UAE’s Falcon was trained on 384 A100 chips over two months earlier this year. “I was extremely impressed by the model, considering the resources they used. For a while, it was among the best models in the open-source world,” said one leading AI researcher.