Eliyan Raises $145 Million for AI Interconnects
Eliyan raised a $145 million Series C round at a $1 billion valuation on July 29, 2026, according to SiliconANGLE and GamesBeat. Seligman Ventures led the financing, with Cisco and Lumentum participating. The Silicon Valley startup develops SERDES interconnects and chiplets intended to link compute, memory, and networking across AI infrastructure.
Eliyan raised $145 million in a Series C financing at a $1 billion valuation, according to SiliconANGLE and GamesBeat. Seligman Ventures led the round. SiliconANGLE named Cisco Ventures and Lumentum as backers, while GamesBeat referred to Cisco Investments and Lumentum as new strategic investors.
GamesBeat reports that Umesh Padval, managing partner of Seligman Ventures, is joining Eliyan's board as part of the financing. The publication also reported that Patrick Soheili, Eliyan cofounder, described the funding as supporting expansion into electro-optical interconnect technology for AI infrastructure.
Connectivity across AI systems
Eliyan develops SERDES interconnect chips and chiplets for die-to-die, chip-to-chip, and rack-to-rack connectivity. GamesBeat describes the technology as addressing connectivity requirements spanning package-level integration, multi-package scale-up systems, memory expansion, and high-bandwidth infrastructure fabrics.
SiliconANGLE identifies Eliyan's products as NuLink PHYs and NuGear chiplets. The outlet reports that the products are intended to link compute, memory, and networking while increasing bandwidth capacity and power efficiency in AI clusters.
The funding arrives as AI infrastructure operators are building larger accelerator clusters for training and inference. According to SiliconANGLE, scaling those clusters increases power and bandwidth demands, and constrained data movement can leave compute resources waiting for input data rather than executing workloads.
Why interconnects matter
In comparable large-scale AI systems, faster accelerators do not by themselves resolve bottlenecks caused by memory access, package-level links, or network fabric capacity. Training and distributed inference performance depend on the interaction among accelerator compute, memory bandwidth, network topology, link reliability, and power delivery.
Eliyan's financing is therefore notable for infrastructure practitioners because it targets a layer beneath model software: the physical and electrical pathways that determine how efficiently systems exchange data. Public reporting does not provide benchmark results, customer deployments, or production-volume details, leaving the practical performance and adoption of the company's components open questions.
Key Points
- 1Eliyan raised $145 million at a $1 billion valuation, bringing substantial new capital to AI interconnect hardware development.
- 2Its SERDES and chiplet products target data movement across packages, chips, and racks, where bandwidth and power constraints affect cluster utilization.
- 3Comparable AI cluster designs increasingly depend on interconnect, memory, and power characteristics alongside accelerator throughput for end-to-end performance.
Scoring Rationale
The $145 million round and $1 billion valuation make this a notable AI infrastructure financing event. Interconnect bandwidth, reliability, and power efficiency are increasingly relevant to teams designing large accelerator clusters, although Eliyan has not disclosed public performance benchmarks or customer deployment details in the cited coverage.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
