AMD and Cerebras team up on ultra low-latency AI systems

AMD CEO Lisa Su and Cerebras CEO Andrew Feldman announced a partnership to develop ultra low-latency AI systems, combining AMD's traditional AI chips with Cerebras' wafer-scale integration technology. The first products will debut later this year in the Cerebras cloud, targeting competition with Nvidia and Intel.

Become a member of GB MAX to gain exclusive access to the industry and to the most influential global B2B leadership community in the business of gaming, entertainment, and tech. Join now https://go.gamesbeat.com/gb-max/ and also get a VIP ticket to GamesBeat Next Nov 2-3, SF .AMD CEO Lisa Su and Cerebras CEO Andrew Feldman announced on stage that they have teamed up in a partnership. The two companies are rivals, but they also have common rivals in the AI chip business in the form of Nvidia and Intel. “We saw a partnership that could extend our capabilities and our footprint and there was no better partner than AMD,” Feldman said. AMD makes AI chips that are traditional semiconductor chips that are sliced from wafers, and Su announced new generations. But Cerebras has wafer-scale integration, meaning its chips are all part of a slice of a 12-inch wafer, with hundreds of CPUs on a single slice of silicon, all networked together. It’s targeted at the ultra low-latency segment. The first products will debut together later this year in the Cerebras cloud, Feldman said. “We can’t wait to get this in” custom hands, Su said. Su and Feldman know each other well, as AMD bought Feldman’s previous startup, Sea Micro.