Credit By: PR NEWSWIRE
The “world’s largest supercomputer” has been unveiled by G42, an artificial intelligence and cloud computing startup based in Abu Dhabi, to leverage AI to tackle problems in healthcare, energy, and climate change.
The two companies announced on Thursday that the company has partnered with US-based artificial intelligence company Cerebras Systems to introduce Condor Galaxy, a network of nine connected supercomputers that aims to drastically cut down the time required to train AI models.
Condor Galaxy 1 (CG-1) is the first AI supercomputer on the network, with 54 million cores and a 4 exaflop processing power. A supercomputer’s performance is measured in exaflops.
Two further supercomputers, the CG-2 and CG-3, will be released in the US in early 2024 by Cerebras and G42.
According to Andrew Feldman, CEO of Cerebras Systems, the Frontier supercomputer at the Oak Ridge National Laboratory of the US Department of Energy was previously the largest supercomputer in the world, with a performance of 1.1 exaflops.
By 2030, generative AI could bring the GCC countries $23.5 billion in economic advantages.
Google and Nvidia collaborate to train AI on A3 supercomputers.
According to Talal Alkaissi, CEO of G42 Cloud, a division of G42, “the UAE has a national AI plan, part of which is to use AI to improve productivity at all levels across the government and economy.”
And this partnership’s computing will be very important in enabling disruptive change in various industries.
According to him, G42 is already developing AI applications for various markets, including big data analytics, sports, space exploration, smart cities, smart mobility, energy transition, and health care.
The announcement establishes the infrastructure to support crucial research, provide applications, and enhance business and personal quality of life.
According to Mr. Feldman, Candour Galaxy 1 has started operating in Santa Clara, California, and G42 scientists and researchers have already put cutting-edge AI models into operation.
“And together in partnership, we plan to expand to three interconnected US-based AI supercomputers and then 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 continued.
Supercomputers are much more potent than general-purpose computers. They are frequently employed to tackle the most challenging issues in the world, including the creation of pharmaceuticals, the discovery of oil and gas reserves, and the forecasting of the weather, among others.
At the I/O conference in Mountain View, California, Google also unveiled a new A3 supercomputer in May that would be used to train AI and machine learning models.
The new technology was developed in collaboration with Nvidia, an AI hardware and software leader that creates graphics processing units for various markets. It aims to provide a full range of GPU (graphics processing unit) options for the training and inference of machine learning models.
GPUs are advantageous for machine learning, video editing, and gaming applications because they can simultaneously process multiple jobs.
Elon Musk, a multibillionaire, also puts $1 billion into his Dojo supercomputer.
Customers can benefit from the performance of an AI supercomputer using the CG-1 cloud service from Cerebras and G42 without having to manage or distribute models across physical systems. It is intended to make it simple and rapid for G42 and its cloud users to train substantial, ground-breaking models.
According to the firms, the Cerebras-G42 alliance has already advanced cutting-edge AI models in Arabic bilingual chat, healthcare, and climate studies.
According to Brainy Insights, the global generative AI industry is expected to increase from $8.65 billion in 2017 to $188.62 billion in 2032, expanding at more than 36% annually.
According to a report released last year by Alibaba Damo Academy, the international research arm of China’s Alibaba Group, AI will be a recurring theme in the top 10 technology developments over the next few years. This is predicted to accelerate discoveries across key economic sectors and society.
Mubadala Investment Company supports G42, conducting high-level fundamental and practical research into AI and creating cloud computing for the most difficult use cases. Nine enterprises with 22,000 people are part of the group.
Additionally, the firm received a “substantial” investment from US-based private equity management Silver Lake. As part of the agreement, Egon Durban, the managing partner and co-chief executive of Silver Lake, joined the board of directors of G42 in 2021.
Numerous cloud providers have announced enormous GPU clusters that cost billions of dollars to create but are very challenging to operate, according to Mr Feldman.
“It takes months of work from dozens of individuals with specialized knowledge to distribute a single model among thousands of small GPUs. The CG-1 solves this problem. One person can set up a generative AI model in minutes rather than months.
According to Mr. Feldman, the adoption of AI across various business processes is exploding across the globe.
“This is not a US phenomenon, which is the true story here. Cerebras and G42 have formed a strategic relationship to participate in this revolution, a worldwide phenomenon that may be the biggest technological upheaval in a generation.