Sunnyvale, California AI-compute company founded 2015; builds wafer-scale WSE-3/CS-3 chips and, since August 2026, the CS-4 accelerator.
Cerebras Systems is an AI compute company headquartered in Sunnyvale, California, founded in 2015. The company designs wafer-scale processors and systems built for AI training and inference, positioned as an alternative to GPU clusters for running large language models.
Its Wafer-Scale Engine 3 (WSE-3) is fabricated on TSMC's 5-nanometer process and carries 4 trillion transistors and 900,000 AI-optimized cores on a single 46,225 mm2 wafer, with 44 GB of on-chip SRAM and 21 PB/s of memory bandwidth. The WSE-3 powers the CS-3 system, a water-cooled unit that Cerebras states can train models with up to 24 trillion parameters and, combined with external MemoryX memory and SwarmX interconnect hardware, scales to clusters of thousands of chips. In August 2026 Cerebras introduced the CS-4 system as the successor to CS-3, describing it as the fastest AI inference accelerator it has built.
Cerebras also runs a cloud inference service offering OpenAI-API-compatible access to open models such as Llama, Qwen, and GPT-OSS, alongside dedicated private-cloud and on-premises deployment options. Customers and partners named on its site include OpenAI, AWS, Meta, GSK, Mayo Clinic, Notion, AlphaSense, Cognition, and Argonne National Laboratory, spanning agentic coding, enterprise search, drug discovery, and scientific computing workloads.
Primary type & automation activities this supplier delivers:
Largest chip ever built: 4 trillion transistors and 900,000 AI cores on a single wafer.
Wafer-scale AI system training models up to 24 trillion parameters, scalable to 2,048-system clusters.
OpenAI-API-compatible cloud serving open frontier models at world-record token speeds.
Contact Cerebras Systems
WEBSITE
https://www.cerebras.ai/HEADQUARTERS
United States
Company Facts
Founded
2015
Primary Role
Component Supplier
Company Size
-
Primary Region
North America
Annual Sales
-
Funding Stage
-
Funding Total
-