|
Listen to this story
|
The American AI company Cerebras Systems and G42, an Abu Dhabi-based AI holding company, have announced the development of Condor Galaxy 3 (CG-3), the latest addition to their AI supercomputing constellation. It features 64 of Cerebras’ newly launched CS-3 systems, all powered by the cutting-edge Wafer-Scale Engine 3 (WSE-3), CG-3 is set to deliver 8 exaFLOPs of AI computing power. It will be available in Q2 2024.
“The existing Condor Galaxy network has trained some of the leading open-source models in the industry, with tens of thousands of downloads. By doubling the capacity to 16 exaFLOPs, we look forward to seeing the next wave of innovation Condor Galaxy supercomputers can enable,” said Kiril Evtimov, Group CTO of G42.
CS-3 is the latest AI chip with 4 trillion transistors that was also unveiled by Cerebras Systems. The WSE-3 delivers twice the performance at the same power and for the same price as the previous generation part, where the WSE-3 delivers 125 petaflops of peak AI performance per chip.It has been designed for training the industry’s largest AI models.
The latest announcement comes after last November when the company announced the 2nd phase of the supercomputer called Condor Galaxy 2. Condor Galaxy has been powering generative AI models including the Arabic model Jais-30B. The latest CG-3 is only going to further push AI supercomputing innovations.
G42’s Rapid Expansion
While this announcement marks their 3rd supercomputer, the company has been spearheading the AI race with its hardware. Last year, the company partnered with OpenAI. The company also partnered with Microsoft to boost cloud and technology infrastructure in the UAE. G42 is also expanding in India by building a team that is spearheaded by former Xiaomi head Manu Kumar Jain.
