Moving data between AI chips has become one of the computing industry’s most expensive bottlenecks, and a small Santa Clara startup just raised $145 million to fix it. Eliyan announced its Series C round on Wednesday at a $1 billion valuation, earning unicorn status just under five years after founding.
The round was led by Seligman Ventures, with participation from Cisco Systems, Lumentum, and former Mellanox board member Umesh Padval. Eliyan develops high-speed interconnect technology and optical chiplets that help AI processors exchange data more efficiently, a problem the company says leaves 60 to 70 percent of GPU capacity idle while chips wait for data to arrive.
“People have tried to build faster compute, faster processors and GPUs,” CEO Ramin Farjadrad told Reuters. “Today we are at the point that maybe only 30 percent to 40 percent of the GPUs are being used, mainly because of how fast they can receive the data to process.”
Eliyan licenses its interconnect technology to semiconductor companies rather than building its own AI processors, positioning itself as an independent alternative to Nvidia’s Mellanox ecosystem and Broadcom-based solutions from cloud providers. Initial chiplet product shipments are expected this year, with revenue projected to grow from low millions in 2025 to hundreds of millions by 2027 as commercial deployments accelerate.
