Cerebras

Leader

Cerebras designs wafer-scale AI accelerators and integrated systems and provides a cloud inference API to run AI training and high-throughput inference workloads by placing large compute and SRAM on a single wafer-scale chip to reduce multi-chip communication overhead.