Cerebras
Headquarters
Sunnyvale, CAFounded
2015CEO / Founder
Andrew FeldmanEmployees
501-1000Funding
Public — $720MStock
CBRSProducts
1 product listedAvg. AI Score
8.0/10Website
cerebras.ai/
Profile updated
About Cerebras
Cerebras Systems is a semiconductor and AI computing company founded in 2016 by Andrew Feldman, Gary Lauterbach, and others, and headquartered in Sunnyvale, California. The company focuses on building specialized hardware and software for AI workloads, particularly large-scale neural network training and inference.
Cerebras's primary product is the Wafer-Scale Engine (WSE), a processor built on a single silicon wafer rather than individual dies, giving it significantly more cores and on-chip memory than conventional GPUs. The company also offers the CS-2 and CS-3 systems, which house the WSE chips, and provides cloud-based access to its hardware through its Cerebras Cloud platform. Customers have included national laboratories such as Argonne and Lawrence Livermore, as well as pharmaceutical and research organizations running large AI models.
Cerebras raised approximately $720 million across multiple funding rounds, including a $250 million Series F in 2021. The company filed for an IPO in 2024, seeking to go public on the Nasdaq under the ticker CBRS, though the offering faced regulatory scrutiny related to a major customer. As of 2024, the company employs several hundred people and has positioned itself as an alternative to Nvidia's GPU-based AI infrastructure.
Specialties
Financials
Dated, source-cited figures. Estimates are marked.
Products by Cerebras
All products by Cerebras reviewed by our AI panel of experts.
How Our Review Panel Scores Cerebras
Six independent reviewer personalities assessed Cerebras's product on their own terms. They landed within 0.4 points of each other, an unusually strong consensus.
Averaged across 1 reviewed product in our catalog. Scores are out of 10.
Competitive Landscape
Who Cerebras competes with, and how they differ.
The incumbent Cerebras is built to displace. NVIDIA's GPU-plus-CUDA stack holds the overwhelming majority of AI training and inference deployments; Cerebras argues a single wafer-scale chip avoids the multi-GPU interconnect overhead that dominates large-model work.
1 product reviewed on TopReviewed
The closest strategic analog — custom non-GPU silicon (LPU) sold primarily as a high-speed inference cloud rather than as boxes. Both compete on tokens-per-second rather than raw FLOPS.
1 product reviewed on TopReviewedThe other well-funded reconfigurable-dataflow accelerator startup targeting enterprise and sovereign AI deployments, and the most direct rival for the on-premises and government contracts Cerebras pursues.
1 product reviewed on TopReviewed
Competes for the same training and inference budgets from the GPU side, renting NVIDIA capacity at scale. Buyers weighing Cerebras systems typically price CoreWeave capacity as the alternative.
1 product reviewed on TopReviewed
An inference-cloud rival for Cerebras Inference. Runs open models on GPU fleets, competing on price and throughput rather than proprietary silicon.
1 product reviewed on TopReviewed
Another GPU-based inference platform competing directly with Cerebras Inference on latency and cost per token for open-weight model serving.
1 product reviewed on TopReviewedCompany Timeline
Key milestones in Cerebras's history, with sources.
- Launch
CS-4 introduced
Cerebras introduced the CS-4, the fourth generation of its wafer-scale system.
en.wikipedia.org - Milestone
Nasdaq IPO (CBRS) — $5.55B raised at $185 per share
Cerebras sold 30 million shares at $185 in the largest US technology IPO since 2019. The stock opened at $350 and finished its first day up 68%.
en.wikipedia.org - Partnership
OpenAI agreement — $10B for 750 megawatts of compute
OpenAI contracted for roughly 750 megawatts of Cerebras computing capacity through 2028 in a deal reported at $10 billion.
en.wikipedia.org - Funding
Series G — $1.1B at an $8.1B valuation
The round raised $1.1 billion at a reported $8.1 billion valuation while the company remained private.
en.wikipedia.org - Milestone
Filed Form S-1 for a US IPO
First registration statement filed with the SEC under CIK 0002021728. The offering was not completed at the time; Cerebras refiled in April 2026.
sec.gov - Launch
WSE-3 and CS-3 — 4 trillion transistors on 5nm
The third-generation Wafer-Scale Engine carried 900,000 AI-optimized cores and 4 trillion transistors on a 5nm process. Time named it one of the best inventions of 2024.
en.wikipedia.org - Partnership
G42 partnership — $100M+ Condor Galaxy supercomputers
Abu Dhabi-based G42 committed more than $100 million to a series of Cerebras-built Condor Galaxy AI supercomputers (CG-1 through CG-3).
en.wikipedia.org - Funding
Series F — $250M at a $4B+ valuation
Alpha Wave Ventures and the Abu Dhabi Growth Fund led the round, lifting the valuation above $4 billion.
en.wikipedia.org - Launch
WSE-2 and CS-2 — 850,000 cores on TSMC 7nm
The second-generation Wafer-Scale Engine carried 850,000 cores and 2.6 trillion transistors, manufactured by TSMC on a 7nm process.
en.wikipedia.org - Funding
Series E — $270M at a $2.4B valuation
The round valued the company at $2.4 billion, roughly doubling its 2018 Series D mark.
en.wikipedia.org - Launch
WSE-1 and CS-1 unveiled — 1.2 trillion transistors
Cerebras revealed the first-generation Wafer-Scale Engine, a single processor cut from an entire silicon wafer carrying 400,000 AI cores and 1.2 trillion transistors, alongside the CS-1 system.
en.wikipedia.org - Founded
Founded in Sunnyvale by the SeaMicro team
Andrew Feldman, Gary Lauterbach, Michael James, Sean Lie and Jean-Philippe Fricker founded Cerebras to build wafer-scale AI accelerators. The five had previously built SeaMicro, which AMD acquired for $334 million in 2012.
en.wikipedia.org
Explore AI Software Reviews
Browse multi-perspective AI panel reviews across hundreds of AI tools, agents, and platforms. Find the right software with insights from CTO, Developer, Marketer, Finance, and User perspectives.