2.6 Trillion Transistors One company's quest to build the world's largest computer chip

Andrew Feldman, CEO of Cerebras Systems with the Cerebras Wafer Scale Engine (WSE-2) - the largest chip ever built to accelerate artificial intelligence work by orders of magnitude beyond the current state of the art.

Scroll this

It seems like every week there’s a new story about how artificial intelligence (AI) is leading to breakthroughs in everything from drug discovery to saving the honeybees.

What’s not often talked about is how these breakthroughs depend on staggering amounts of computing power. So much so that tech giants such as Apple, Amazon and Google are in a race to design the next generation of computer chips tuned specifically for the needs of AI.

While it’s too early to declare a winner, a startup called Cerebras has beaten them to the punch by developing the world’s largest AI computer chip.

Which brings us to a recent assignment to photograph Cerebras CEO Andrew Feldman and his chip for IEEE Spectrum

First off, this isn’t your ordinary computer chip. The Cerebras chip is the size of a dinner plate and packed with a mind blowing 2.6 trillion transistors. Their idea was to create a chip so large that all the data needed by an AI can stay on the chip itself instead of slowly being shuttled back and forth to the computer’s memory. 

Photographing this chip was a fun challenge as it‘s razor thin and mirror-like which meant that we were in full reflection wrangling mode that day taping duvetyne all over the place.

And special thanks to Andrew for being such a willing subject and even agreeing to lay on the floor to get the right angle for one of these shots. Not all CEO’s are willing to go the extra mile like that. 🙏🙏🙏

Submit a comment

Your email address will not be published. Required fields are marked *