On August 18, 2026, Cerebras unveiled the CS-4, a fourth-generation AI accelerator built on three new Wafer Scale Engine 3 Turbo processors under a new modular Nexus rack-scale architecture. The company says the system delivers up to 30 times faster inference than GPU-based systems, offers 10 times higher throughput per watt than its CS-3 predecessor, and can support models exceeding 10 trillion parameters. First CS-4 shipments began in the current quarter.
