AMD and Cerebras Team Up on Ultra-Low-Latency AI Inference, Targeting 5x Efficiency Gain
AMD has more than doubled off its 52-week low of $149.22.
Summary
AMD and Cerebras are partnering to combine AMD Helios rackscale solutions with Cerebras Wafer-Scale Engine into a single disaggregated inference workflow. The joint solution targets ultra-low-latency AI applications, claiming up to 5x higher tokens per second per watt. Cerebras plans to deploy AMD Helios in its data centers, with availability expected through Cerebras Cloud in the second half of 2026. This follows AMD's recent strategic moves in AI, including a partnership with Anthropic for up to 2 gigawatts of AMD capacity announced yesterday. The collaboration strengthens AMD's position in the fast-growing inference market and could drive incremental data center GPU demand.
At the time of this announcement, AMD was trading at $533.76 on NASDAQ in the Technology sector, with a market capitalization of approximately $870.3B. The 52-week trading range was $149.22 to $584.73. This news item was assessed with positive market sentiment and an importance score of 7 out of 10. Source: GlobeNewswire.