CEA, AMD team on next generation sustainable AI hardware
Cette publication existe aussi en Français
French research lab CEA has signed a deal with AMD to develop architectures for the next generation of sustainable AI hardware.
AMD is gaining traction in AI compute in datacentres using its CDNA GPU architecture and Instinct MI300 AI inference cards. The Instinct MI350 series, with the CDNA 4 architecture, is expected to be available in 2025 bringing up to a 35x increase in AI inference performance compared to previous Instinct MI300 Series cards. The next generation MI400 is already in development with shipping planned in 2026 using the next generation CDNA architecture.
The letter of intent with the Commissariat à l’énergie atomique et aux énergies alternatives (CEA) will focus on more energy-efficient systems for the most compute-intensive AI workloads.
The two are planning to hold a European symposium on the future of AI compute later this year with European stakeholders and global technology providers, startups, supercomputing centres, universities and policy makers to work on the challenge.
Last year AMD bought Europe’s largest AI lab, Silo.ai, to boost its development. It also has a deal with Fujitsu in Japan to combine its Monaka 2nm ARM chip with AMD’s Instinct AI accelerator boards.
“AI computing continues to drive innovation across industries, and international collaboration is critical to pushing the boundaries of what’s possible,” said Ralph Wittig, Corporate Fellow and head of research, AMD. “Through this collaboration with CEA and leading French engineers, we aim to bring cutting-edge AI research closer to real-world applications by advancing system architectures that meet the demands of tomorrow’s AI workloads, while growing the joint research and development opportunities between the U.S. and France.”
“CEA is committed to driving innovation in AI computing by advancing next-generation technologies opening the road for disruptive architectures that balance performance and energy efficiency. Our collaboration with AMD represents a significant step toward fostering international cooperation in high-performance computing, bringing together world-class expertise to address the growing demands of AI workloads,” said Julie Galland, Director of the Technological Research division at CEA, “By combining CEA’s research leadership with AMD’s cutting-edge technology, we aim to develop breakthrough solutions that will shape the future of AI computing in Europe and beyond.”
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
