The companies say their systems can accelerate separate stages of AI workloads, from prompt processing to token generation.
AMD and Cerebras partner to distribute AI inference workloads via Cerebras Cloud, with Cerebras deploying AMD Helios systems. Move reflects shift to heterogeneous chips per AI stage, following Anthropic's AMD adoption as computing market targets $2 trillion by 2030.
Add Axios as your preferred source to
see more of our stories on Google.
Lisa Su, CEO of AMD. Photo: Bridget Bennett/Bloomberg via Getty Images
AMD on Thursday announced a partnership with Cerebras that will let customers split AI inference workloads across the companies' systems.
Why it matters: Inference is the part of AI computing that turns a trained model into a response, making it central to the speed and cost of everyday AI services.
The enemy of my enemy is my friend
Helios and Cerebras go up against Nvidia's NVL72 and Groq LPX
AMD is betting the future of AI won't rely on a single chip, unveiling a Cerebras partnership and new Helios system to challenge…
AMD and Cerebras on Wednesday separately announced partnerships related to their AI chips. AMD stock and Cerebras stock both rose.
Cerebras and AMD agreed to work together, pairing their technology for AI systems.
AMD expects to make “tens of billions of dollars.”
