Security· August 19, 2026 at 12:00 a.m.
Cerebras Unveils Next-Gen AI Accelerators Doubling Performance
Key takeaways
- Cerebras WSE-3T promises double the compute, memory fabric, and I/O bandwidth
- The new chip runs at approximately 2.8 GHz, up from 1.4 GHz in the previous generation
- Each WSE-3T boasts 250 petaFLOPS of AI compute, 44 GB of SRAM, good for 43.2 PB/s of memory bandwidth, and 2.4 Tbps of off-die connectivity
Cerebras Systems, a chipmaker specializing in AI accelerators, has announced its next-generation Wafer Scale Engine (WSE) and Nexus rack systems. The new WSE-3T promises to double the compute, memory fabric, and I/O bandwidth of the previous generation while maintaining the same process tech, wafer area size, transistor count, core count, and SRAM capacity. This is achieved by pushing the existing wafer scale engine harder, with a primary focus on power delivery efficiency.
