AMD, CEREBRAS PARTNER ON AI INFERENCE CHIP

The partnership integrates AMD's server architecture with Cerebras' custom AI processors to create a unified inference solution. AMD's Helios platform will connect to Cerebras' wafer-scale chip technology, designed to handle large language model deployments more efficiently than traditional GPU setups. Cerebras wafer-scale processors feature a single chip with millions of cores, eliminating traditional interconnect bottlenecks between multiple processors. The collaboration positions both companies to compete in the competitive AI inference market, where demand continues to grow as enterprises deploy generative AI applications. AMD has been expanding its AI portfolio beyond graphics processors, while Cerebras has focused specifically on large-scale AI workloads. The partnership leverages complementary strengths: AMD's data center experience and Cerebras' specialized chip architecture. Details on pricing, availability, and performance metrics were not disclosed.