Cerebras Systems
Cerebras Systems is an American AI chip design company focused on providing high-performance computing solutions for large-scale AI model training and inference. Its core product uses a proprietary wafer-scale chip architecture (Wafer Scale Engine, WSE), integrating an entire silicon wafer into a single processor to achieve an extremely high number of compute cores and on-chip memory bandwidth. The company targets cloud service providers, research institutions, and enterprise AI application scenarios.
Core Facts
Cerebras' latest-generation WSE-3 chip has an area of approximately 46,225 square millimeters, integrates 4 trillion transistors and 900,000 AI-optimized cores, and is more than 50 times the area of NVIDIA's H100 GPU.
Cerebras Systems focuses on wafer-scale chips (Wafer-Scale Engine), fabricating an entire silicon wafer into a single processor
Timeline (last 90 days)
Cerebras Systems co-founder and CEO Andrew Feldman will discuss the topic of AI hardware scaling limits at TechCrunch Disrupt 2026.
Andrew Feldman is the co-founder and CEO of Cerebras Systems.
Cerebras aims to redefine the efficiency and density of compute units at the chip architecture level in order to achieve higher effective compute within the same energy and space budget.
All Facts (5)
Cerebras' latest-generation WSE-3 chip has an area of approximately 46,225 square millimeters, integrates 4 trillion transistors and 900,000 AI-optimized cores, and is more than 50 times the area of NVIDIA's H100 GPU.
80%VerifiedCerebras Systems focuses on wafer-scale chips (Wafer-Scale Engine), fabricating an entire silicon wafer into a single processor
75%UnverifiedAndrew Feldman is the co-founder and CEO of Cerebras Systems.
50%UnverifiedCerebras aims to redefine the efficiency and density of compute units at the chip architecture level in order to achieve higher effective compute within the same energy and space budget.
50%UnverifiedCerebras的WSE（Wafer Scale Engine）将整张硅晶圆作为单一芯片使用，消除芯片间通信延迟，特别适合自回归大语言模型推理
50%