January 28, 2026 lands quietly on the calendar, but nothing about Eliyan Corporation is subtle. Santa Clara born, April 2021 official, 2016 in the lab, this company did not show up to decorate the AI boom. It showed up to deal with the physics. When bandwidth chokes, latency lies, and power budgets tap out, Eliyan steps into the room like it owns the wiring behind the walls.
This story starts with Ramin Farjadrad, co-founder and Chief Executive Officer, a Stanford-trained engineer with more than 130 patents and a résumé that reads like a semiconductor speedrun. Velio Communications. Aquantia. Marvell. The common thread is not logos, it is interconnect. When chiplets became inevitable, Ramin Farjadrad did not debate it on panels. He built the answer and spun Eliyan out to finish the job.
Eliyan just secured $50 million in strategic funding, bringing total capital raised north of $150 million. The investor list reads like the companies that ship the future rather than comment on it. Samsung Catalyst Fund and Intel Capital returned. AMD, Arm, Meta, Coherent joined. These are not tourists. These are operators who understand that AI systems do not scale on hype. They scale on clean, fast, low power connections that do not blink under load.
The product is where the wordplay earns its keep. NuLink is not about shouting faster, it is about listening smarter. Sixty four gigabits per second per bump on 3 nanometer silicon, standard packaging, five terabits per second per millimeter, and power efficiency that cuts alternatives in half. Simultaneous bidirectional signaling lets one wire talk and listen at the same time, doubling throughput without doubling heat. The industry noticed. Open Compute Project adopted it as Bunch of Wires. UCIe compatibility keeps it fluent across ecosystems.
Patrick Soheili, co-founder and Head of Business and Corporate Development, translates this physics into momentum. Syrus Ziai, co-founder and Vice President of Engineering, turns it into silicon that tapes out on 5 nanometer and 3 nanometer without excuses. Design centers now stretch from Santa Clara to Eindhoven, Frankfurt, Toronto, and Vancouver, because chiplet economics are global whether policy likes it or not.
This $50 million is not about survival. It is about qualification, deployment, and scale as AI infrastructure strains under its own ambition. When hyperscalers, memory giants, and CPU architects all lean in at once, they are telling you where the bottleneck really is. Follow the wires. Watch who controls them. The next phase of AI does not need louder promises. It needs Eliyan to keep the signal clean while everything else gets noisy.
