Cerebras Brand Guidelines: The Sunnyvale Maker of a Chip the Size of a Whole Wafer

A team that had built servers before.

Cerebras Systems was founded in 2015, the company says on its own history page, by Andrew Feldman, Gary Lauterbach, Michael James, Sean Lie and Jean-Philippe Fricker. English Wikipedia gives the same year and the same five names, and notes that they had worked together at SeaMicro, a server maker started by Feldman and Lauterbach in 2007 and sold to AMD in 2012. Crunchbase News reported in May 2026 that the venture firms Benchmark, Foundation Capital and Eclipse led the company’s first funding round in 2016. The company is based in Sunnyvale, California, and English Wikipedia lists further offices in San Diego, Toronto and Bangalore.

The founding idea, as the company describes it, was to bring wafer-scale computing to market. A normal chip is one of many small rectangles cut from a round silicon wafer; Cerebras set out to use almost the whole wafer as one processor. The company says that at the time it was unclear whether such a thing was technically possible. English Wikipedia describes the problems the team had to solve along the way: cooling a very large piece of silicon, routing around the small manufacturing defects every wafer contains, and packaging the result so that it could receive power and data. It records that a working product arrived in July 2019, after years of trial and error.

That product was announced in August 2019 as the first Wafer-Scale Engine, or WSE-1, together with the CS-1, a rack-mounted system built around a single wafer. English Wikipedia says early customers were research institutions and life-sciences companies, among them national laboratories in the United States and the Edinburgh Parallel Computing Centre in Scotland.

The biggest chip, put on display.

Each generation of the Cerebras chip has been introduced with a matching system. The WSE-2 and CS-2 followed in April 2021; English Wikipedia says the chip was made on TSMC’s 7-nanometre process. The WSE-3, on a 5-nanometre process, and the CS-3 followed in March 2024. English Wikipedia gives the WSE-3 a size of 215 millimetres on each side and says it takes up an entire wafer; it adds that the chips are manufactured by TSMC, which it describes as the only company currently able to make them. In August 2026, the company says, it launched the CS-4, a system built around three processors it calls WSE-3 Turbo.

The chip’s size has also been put on display. In August 2022, English Wikipedia records, the Computer History Museum in Mountain View opened a display of the WSE-2 titled “The Biggest Chip In the World.” Time named the CS-3 to its list of the best inventions of 2024, and in May 2024 put Cerebras on its list of the 100 most influential companies, according to the same article. The company also cites recognition from HPCwire’s readers and editors in 2025 and from Fast Company’s Most Innovative Companies list in 2026.

The wafers have also been grouped into larger machines. English Wikipedia describes the Andromeda cluster of sixteen WSE-2 systems, unveiled in November 2022, and the Condor Galaxy supercomputers built with the Abu Dhabi technology group G42 from July 2023 onward. In 2022, it adds, researchers at Argonne National Laboratory won the Gordon Bell Special Prize for COVID-19 research using CS-2 systems alongside other hardware.

From August 2024 the company also sold access to its chips as a cloud inference service, so that developers could use the hardware without buying a system. English Wikipedia records that it announced plans for six new data centres in North America and France in March 2025, that Meta agreed to use Cerebras for its Llama API in April 2025, and that Cerebras signed agreements with OpenAI in January 2026 and with Amazon Web Services in March 2026.

How AI chip makers show themselves.

AI chip companies sell something most of their customers will never see. The processor sits inside a server, the server sits in a rack, and the rack sits in a data hall far from the people who use the models running on it. So the industry’s visual language has grown up around making that hidden hardware visible: a whole silicon wafer, as wide as a dinner plate, held up on a keynote stage; a single chip pinched between finger and thumb for the cameras; a die shot recoloured in neon false colour so that its blocks of memory and compute read like a city map at night.

The second set of images is the data centre. Press photographs tend to show long rows of racks in dark halls, lit by the glow of status lights and cold aisles, with cabling arranged as neatly as a diagram. These pictures stand in for scale, which is what buyers of AI compute are being asked to imagine, and they appear again and again in launch announcements, partner news and investor material.

Keynote slides do much of the remaining work. Model names, token counts and benchmark charts appear beside the chip’s own product name, and the companies’ identities are usually kept spare to let those numbers and images lead: black or white grounds, a single accent colour, a sans-serif wordmark and very little ornament. Several firms have adopted coined product categories, names for their processors that set them apart from the graphics processor.

The shop window is the technical conference. Hot Chips, NeurIPS and the AI Infra Summit are among the events where architectures are presented, demonstrations run and booths compete for engineers’ attention. This guide looks at how each company in the group presents itself, and keeps to what can be checked about who founded it, who owns it and what it makes.

A long road to Nasdaq.

Cerebras’s route to the stock market took nearly two years. The company first filed publicly for an initial public offering in September 2024, Crunchbase News reported. Reuters reported in March 2025 that the listing had been held up by a review by the Committee on Foreign Investment in the United States of an investment in Cerebras by G42; in March 2025, the company later said, the committee cleared that sale of shares. On 3 October 2025 Cerebras said it was withdrawing the planned offering, the Taiwanese site TechNews reported the next day, adding that its chief executive considered the year-old prospectus out of date and that the company still hoped to list. It filed a new registration in April 2026, Futuriom reported.

The shares began trading on the Nasdaq on 14 May 2026 under the ticker CBRS, Crunchbase News and Futuriom reported that day, and Cerebras’s own history page marks the month as the one in which it went public. The company now describes itself as an AI inference and training platform, sold both as on-premise supercomputers and as pay-as-you-go cloud services.

For a design guide, Cerebras shows how a single physical object — a processor as large as the wafer it is cut from — can carry a technology brand, in museum displays and on stage. Compare Nvidia, the graphics-processor maker, and TSMC, the foundry that makes the wafers. For this guide we use black, graphite and a kiln orange with a sans, an unofficial interpretation; no product or logo is reproduced.

Keep Cerebras’s identity on hand — save it to a board in the Made Good inspiration library.