Eliyan Corporation, a Santa Clara, CA based company specializing in connectivity products and technologies for AI and computing systems, has raised $50 million in strategic funding, led by Samsung Catalyst Fund, Intel Capital, and players across the AI and compute ecosystem, including AMD, Arm, Coherent, and Meta.
The involvement of these strategic partners demonstrates strong confidence in Eliyan’s goal of delivering scalable, energy-efficient computing for the next generation of AI and advanced workloads. As AI systems grow beyond single chips and modules, they need high-performance connections both within a package and across chips, systems, and racks.
Read More:Adaptive6 Raises $28M in Series A Funding
The company plans to use the funds to accelerate the production and testing of its interconnect IP and chiplet products, expand partnerships, and help roll out these technologies across AI systems, computing, and edge devices.
This funding is an essential milestone as Eliyan moves toward bringing its NuLink™ PHY and NuGear™ chiplet products to market. These technologies are built to break through memory and I/O limits that hold back scale, power efficiency, and cost in next-generation AI systems.
The NuLink PHY portfolio covers several types of connectivity, including:
The silicon-proven NuLink™ D2D at 64G has been deployed and validated over two years ahead of the broader industry, supporting next-generation memory interconnects. This includes the new SPHBM4e standard, recently announced by JEDEC, which supports next-generation HBM5 bandwidths on standard packaging, highlighting the company’s ability to deliver solutions ahead of industry standards.
Next-generation NuLink™ C2C includes 32G–64Gbps single-ended (NuLink-XS) and 224Gbps differential SerDes, with future 448Gbps (NuLink-XD) SerDes technologies. These are designed to connect multiple packages or modules across boards and systems. The NuLink-X family lets Eliyan support large-scale, disaggregated AI systems with very high bandwidth and about twice the energy efficiency of other solutions.
NuGear™ chiplets are designed to deliver high-performance network connectivity, with link speeds from 1.6T to 12.8Tbps. They target demanding AI accelerators and memory systems, where latency, power, and reliability are critical, and can be combined with optical engines for even better performance.
“This backing from leading companies running AI at massive scale highlights the value of Eliyan’s innovation,” said Ramin Farjadrad, CEO and Co-Founder of Eliyan. “Our silicon proven 64G D2D technology enables efficient on-package chiplet integration, and our roadmap for 224G and 448G chip-to-chip interconnects, along with high-speed 1.6T–12.8T scale-up connectivity and early SPHBM4e silicon extending toward future SHHBM5 positions Eliyan to power the next generation of AI system architectures.”
“As AI systems evolve, they rely more on scalable, energy efficient designs and strong ecosystem collaboration,” said Mohamed Awad, Executive VP of Arm’s Cloud AI Business Unit. “Arm is committed to providing the technologies and partnerships needed to support next generation AI infrastructure at scale.”
“AI infrastructure is expanding rapidly, driving the need for high-performance connectivity to scale compute and memory,” said Dede Goldschmidt, Senior VP and Head of Samsung Catalyst Fund. “We’re excited by Eliyan’s progress since leading their 2024 funding round. The top investors in this new round show the team’s strong execution in both PHY and chiplet technologies.”
About Eliyan
Founded in 2021 and led by CEO Ramin Farjadrad, Eliyan Corporation is driving the chiplet revolution with the industry’s most efficient, scalable interconnect solutions for multi-die architectures. Eliyan’s NuLink™ PHY and chiplet products deliver high bandwidth and low power across both advanced and standard packaging, enabling chipmakers and system companies to boost core counts, memory capacity, and performance while reducing costs and energy use.
Read More:Slice Global Equity Raises $25M in Series A Funding
