Key facts

Founded
2016
HQ
Sunnyvale, United States
Employees
708
Funding stage
Public · Nasdaq: CBRS
Further facts
  • Offices: Cerebras's principal executive offices are at 1237 E. Arques Avenue, Sunnyvale, California.
  • Capital: Cerebras closed its IPO on 15 May 2026, selling 34,500,000 Class A shares at US$185 each, including the underwriters' full option, after raising US$1.0 billion of Series H preferred stock in January 2026.
  • Filing: Cerebras reported revenue of US$180.1 million for the quarter to 30 June 2026, of which US$126.0 million came from cloud and other services.
  • Customers: In January 2026 Cerebras signed a multi-year deal with OpenAI valued at more than US$20 billion to deploy 750 megawatts of Cerebras compute.
  • Leadership: Co-founder Andrew D. Feldman is chief executive and president, Sean Lie is chief technology officer and Robert Komin is chief financial officer.

Focus areas

AI chips LLM inference AI infrastructure AI data centres Model training

About

Cerebras Systems designs and builds AI compute around the Wafer-Scale Engine, a processor made from an entire silicon wafer. Customers buy Cerebras AI supercomputers for their own data centres or use its compute through Cerebras Cloud and partner marketplaces. Its customers include hyperscalers, foundation model labs, AI-native businesses, Fortune 500 companies and sovereign AI initiatives, with OpenAI, G42, MBZUAI and AWS among the largest. Founded in 2016, Cerebras listed on Nasdaq in May 2026 and had 708 people at the end of 2025.

How they use AI

Cerebras builds the hardware and software that run AI models rather than applying AI to another business. The third-generation WSE-3 has 4 trillion transistors and 44 gigabytes of on-chip memory, and each chip sits in a CS-3 system with its own cooling, power delivery and interconnect. Multiple CS-3 systems connect into AI supercomputers deployed on premises or in the cloud. Cerebras Inference serves open-source and partner models at high speed through Cerebras Cloud and partners such as AWS Marketplace, Microsoft Marketplace, OpenRouter and Hugging Face. The Cerebras Compiler compiles PyTorch models directly to the WSE without CUDA or distributed programming. The CS-4 system, launched in August 2026, uses three WSE-3 Turbo processors.

Products and services

  • Wafer-Scale Engine (WSE-3) — Single-wafer AI processor with 4 trillion transistors and 44 GB of on-chip memory
  • CS-3 system — Integrated AI compute system housing the WSE with cooling, power and interconnect
  • CS-4 system — System with three WSE-3 Turbo processors, launched in August 2026
  • Cerebras AI supercomputers — Clusters of Cerebras systems for on-premises training and inference
  • Cerebras Cloud and Inference — Consumption-based access to Cerebras compute and high-speed model inference

Notable customers / partners

Key people

  • Andrew D. Feldman — Co-Founder, Chief Executive Officer and President
  • Robert Komin — Chief Financial Officer
  • Sean Lie — Co-Founder and Chief Technology Officer
  • Dhiraj Mallick — Chief Operating Officer

Investors

  • Tiger Global
  • Benchmark
  • Fidelity Management & Research Company
  • Atreides Management
  • Alpha Wave Global
  • Altimeter
  • AMD
  • Coatue
  • 1789 Capital