Cerebras Partners With AMD and CrowdStrike for High-Speed AI

Over two days in July 2026, Cerebras Systems entered two strategic partnerships to scale its AI inference capabilities and secure its infrastructure. On July 22, the company partnered with CrowdStrike to integrate high-speed AI inference with AI-native security. CrowdStrike is utilizing Cerebras technology to power its Falcon AI Detection and Response models to counter AI-accelerated threats, while Cerebras has standardized its business security on the CrowdStrike Falcon platform.

The following day, Cerebras and AMD unveiled a disaggregated AI inference solution at the Advancing AI 2026 event. This platform combines AMD Helios rackscale solutions, which act as a high-throughput prompt engine, with the Cerebras Wafer-Scale Engine for ultra-low-latency token generation. The companies claim this joint architecture can deliver up to five times higher tokens per second per watt than a Cerebras-only configuration.

This technical collaboration targets latency-sensitive applications including robotics, autonomous agents, and real-time copilots. Cerebras plans to deploy AMD Helios in its data centers, with the combined service expected to be available via Cerebras Cloud in the second half of 2026.