Abu Dhabi's G42 unveils world’s largest supercomputer for AI training

G42, the Abu Dhabi-based artificial intelligence and cloud computing company, has launched the “world’s largest supercomputer” that aims to address challenges in health care, energy and climate action using artificial intelligence.

The company has joined forces with US-based AI firm Cerebras Systems to launch Condor Galaxy, a network of nine interconnected supercomputers, that promises to significantly reduce AI model training time, the two companies said on Thursday.

The first AI supercomputer on the network, Condor Galaxy 1 (CG-1), has 4 exaflops of capacity and 54 million cores. Exaflop is a measure of performance for a supercomputer.

Cerebras and G42 are planning to introduce two more such supercomputers – CG-2 and CG-3 – in the US in early 2024.

The Frontier supercomputer at the US Department of Energy’s Oak Ridge National Laboratory was earlier the world's largest supercomputer, with 1.1 exaflops of performance, said Andrew Feldman, chief executive of Cerebras Systems.

“In the UAE, there's a national AI plan, part of which is to use AI to improve productivity at all levels across the government and economy,” Talal Alkaissi, chief executive of G42 Cloud, a subsidiary of G42, said during an online briefing.

“And we believe the compute we are providing through this partnership will play a large role in the ability to create disruptive change in various different sectors.”

G42 is already working on AI applications across industries including health care, smart cities, smart mobility, energy transition, financial services, sports, space exploration and big data analytics, he said.

“The announcement puts in place the much-needed infrastructure to power the vital research, deliver applications and improve efficiencies in our enterprises and the quality of our lives.”

Mr Feldman said Candor Galaxy 1 has begun operations in Santa Clara, California and G42 scientists researchers have pioneering AI models up and running already.

“And together in partnership, we plan to expand to three interconnected US-based AI supercomputers and then to expand to six more for a total of nine, 4 exaflop AI supercomputers linked together producing a 36 exaflop distributed AI supercomputer, the largest in the world,” he added.

Supercomputers are far more powerful than general-purpose computers and are typically used to address the most demanding problems in the world, including the development of medicines, the exploration of oil and gas reserves and weather forecasts, among others.

In May, Google also announced a new A3 supercomputer to train machine learning and AI models at the I/O conference in Mountain View, California.

Built in partnership with Nvidia – a global leader in AI hardware and software that designs and manufactures graphics processing units for various industries – the new technology aims to offer a complete range of GPU (graphics processing unit) options for the training and inference of machine learning models.

GPUs can process various tasks simultaneously, making them useful for machine learning, video editing and gaming applications.

Billionaire Elon Musk is also investing $1 billion in his Dojo supercomputer.

Cerebras and G42 offer CG-1 as a cloud service, allowing customers to enjoy the performance of an AI supercomputer without having to manage or distribute models over physical systems. It is designed to enable G42 and its cloud customers to train large, ground-breaking models quickly and easily.

The Cerebras-G42 partnership has already advanced state-of-the-art AI models in Arabic bilingual chat, health care and climate studies, the companies said.

The global generative AI market is forecast to reach $188.62 billion by 2032, growing at an annual rate of more than 36 per cent, from $8.65 billion last year, according to Brainy Insights.

AI will be the common theme in the top 10 technology trends in the next few years, and these are expected to accelerate breakthroughs across major economic sectors as well as society, Alibaba Damo Academy, the global research arm of China's Alibaba Group, said in a report last year.

G42, backed by Mubadala Investment Company, is carrying out high-level fundamental and applied research into AI as well as developing cloud computing for the most demanding use cases. The group has nine companies under its umbrella with 22,000 employees.

The company also secured a “substantial” investment from US-based private equity manager Silver Lake, whose managing partner and co-chief executive Egon Durban joined G42's board as part of the deal in 2021.

“Many cloud companies have announced massive GPU clusters that cost billions of dollars to build, but that are extremely difficult to use,” said Mr Feldman said.

“Distributing a single model over thousands of tiny GPUs takes months of time from dozens of people with rare expertise. CG-1 eliminates this challenge. Setting up a generative AI model takes minutes, not months and can be done by a single person.”

The use of AI is surging globally with nearly all geographies adopting it across several business functions, Mr Feldman said.

“The real story here is this is not a US phenomenon. This is a global phenomenon and this is perhaps the largest single technology shift that we've seen in a generation and into this transformation Cerebras and G42 have forged a strategic partnership.”