AMD Partners with Cerebras to Enhance AI Inference Capabilities

AMD has unveiled a strategic partnership with Cerebras aimed at enhancing AI inference capabilities. This collaboration will integrate AMD's Helios GPU rack with Cerebras' wafer-scale AI chip to create an ultra-low latency AI inference platform, which is anticipated to be available through Cerebras Cloud in the second half of 2026.

AMD's CEO, Lisa Su, highlighted that the partnership is focused on delivering high-speed AI inference, specifically targeting enterprise and hyperscale workloads. The company also announced that its new EPYC 'Venice' server processor is in full production and claims a significant performance advantage over Arm-based server chips, asserting a 20% higher performance than Nvidia's CPU offerings.

The Venice processor is expected to start shipping in the fourth quarter of this year, with strong customer demand already noted. This collaboration with Cerebras signifies AMD's ongoing strategy to expand its presence in the AI infrastructure market, moving beyond GPUs to encompass the entire AI stack, from training to inference and enterprise data centers.

In market reactions, AMD shares are currently down 3.06% at $535, while Cerebras shares have risen 4.03% to $218. Technically, AMD's stock has seen a decline but remains above critical moving averages, suggesting a potential bullish outlook if it maintains this position. Conversely, Cerebras is recovering from a significant drop since its IPO, with recent trading showing improved technical indicators.

Overall, this partnership positions AMD to compete more robustly in the AI space, potentially leading to increased sales and profitability as demand for advanced AI solutions grows.