Total 640 words used.
Andrew Feldman is an entrepreneur at his core. He is co-founder and Chief Executive Officer at Cerebras Systems, developer of the world’s fastest AI super computer. Feldman co-founded Cerebras in 2015 with the goal of revolutionizing compute for AI by accelerating computation by up to 1,000 times faster than the current state of the art.
To achieve these goals, Cerebras solved a problem that had remained unsolved in the computer industry for 70 years—how to build, package, and productize wafer scale processors. The Cerebras processor, known as the Wafer Scale Engine, was first of its kind. It is 56 times larger than the second largest chip in existence. Put simply, Cerebras delivered a processor the size of a dinner plate, while the rest of the industry built processors the size of a postage stamp.
Cerebras is now valued at more than $4 billion, having raised more than $720 million in venture financing. Cerebras’ founding team all worked at Feldman’s previous startup, SeaMicro, where the team pioneered the energy efficient server category. SeaMicro, where Feldman was co-founder and CEO, sold to AMD in 2012 for $334 million. More than 50 engineers followed Feldman from SeaMicro and other previous companies to join Cerebras.
2021 was an extraordinary year for Cerebras. In April, the company announced the CS-2 system, the industry’s fastest AI supercomputer. The CS-2 is based on the second-generation Wafer Scale Engine (WSE-2), which is the largest processor ever built. It contains 2.6 trillion transistors whereas the next largest chip contains 45 billion transistors, making the CS-2 2.55 trillion transistors larger than its nearest competitor.
The CS-2 was used to solve previously unsolved problems in AI. Feldman and the Cerebras team also announced the world’s first multi-million core AI cluster and the industry’s only capability to support AI models with up to 120 trillion parameters, approximately the same number of synapses in the human brain. With support for up to 120 trillion parameter models, the world will enter a new era, enabling scientists to train a model in days instead of months, as well as train models that are three orders of magnitude more complex than the largest model today.
Customers have flocked to the Cerebras solution. GlaxoSmithKline (GSK), AstraZeneca, Genentech, AbbVie, and Bayer are all now customers. So too are many of the major supercompute sites in the U.S. and Europe including Argonne National Laboratory, Lawrence Livermore National Laboratory, Pittsburgh Supercomputer Center, and Edinburgh Parallel Computing Centre. Additional customers are found in the US Department of Defense and U.S. intelligence agencies.
Andrew and the Cerebras team secured a Special Mention in the TIME Best Inventions list in November 2021 for the CS-2 system. They were also named by Fast Company as one of the Most Innovative Companies in AI in March 2021.
In November 2021, Feldman secured $250 million in Series F funding for the company, which was led by Alpha Wave Ventures, G42 and Abu Dhabi Growth Fund (ADG). In 2021 Cerebras nearly doubled its size and now has more than 400 engineers across sites in Silicon Valley, San Diego and Toronto Canada as well as sales offices in Tokyo, Japan.
Feldman and the executive team have positioned Cerebras for explosive growth in the next several years, particularly as AI demand continues to increase exponentially. Going forward into 2022, Cerebras is planning to expand its footprint to additional regions, as well as increase its headcount worldwide to more than 600 by the end of next year.
Feldman is a sought-after advisor and currently serves on the board of directors at Natron Energy and on the advisory board of more than a dozen other startups. He is a frequent keynote speaker and guest lecturer at the Stanford Graduate School of Business, his alma mater, where several business cases have been published based on his work.
