Package & SI/PI

Eliyan raises $50 million in funding. Eliyan raised $50 million in a strategic Series C round, bringing total funding to approximately $150 million. Investors include major AI and semiconductor leaders like AMD, Arm, Coherent, Meta, Samsung Catalyst Fund, and Intel Capital, suggesting strong ecosystem support for Eliyan's chiplet interconnect technology amid growing AI demands. Eliyan, based in Santa Clara, California, is a fabless semiconductor company specializing in chiplet interconnect solutions for high performance computing and AI applications. Founded by industry veterans including Ramin Farjadrad (CEO and co-inventor of the Bunch of Wires scheme), Syrus Ziai, and Patrick Soheili, the company addresses limitations in traditional system-on-chip (SoC) designs by enabling efficient multi-die architectures. Their technologies support standards like UCIe and BoW, allowing higher bandwidth, lower power, and reduced costs without relying on advanced packaging like silicon interposers. Eliyan's approach promotes sustainability by minimizing manufacturing waste and power use, targeting sectors like hyperscalers, AI processors, and graphics chips. As of 2024, the company employs around 49 people and holds 22 patents. The $50 million Series C is a strategic investment designed to bolster Eliyan's role in scalable AI systems. It follows a $60 million Series B in March 2024 and a $40 million Series A in November 2022, with no separate disclosed amount for an August 2024 extension noted in some sources. The round emphasizes accelerating manufacturing, qualification, and ecosystem partnerships for deployment in AI infrastructure and edge computing. Investors' involvement highlights collaborative validation: for instance, Meta's participation aligns with hyperscaler needs for efficient computing connectivity. This funding comes amid surging demand for AI hardware, where chiplet architectures are projected to grow significantly, e.g., the high bandwidth memory (HBM) market could expand by 331% in one year alone. Eliyan's solutions, like the NuLink PHY achieving 64Gbps on a 3nm process, could enable up to 4x better performance and half the power of competitors, positioning the company to capture share in a $50 billion chiplet sector. The strategic nature of the round suggests Eliyan is well placed for partnerships, potentially leading to faster adoption in GenAI systems. Eliyan Corporation, headquartered in Santa Clara, California, operates as a fabless semiconductor company dedicated to pioneering chiplet interconnect technologies that facilitate the development of scalable, efficient AI and high performance computing systems. Established by a team of networking and interconnect experts, led by CEO Ramin Farjadrad, who is renowned for inventing the Bunch of Wires (BoW) scheme that has been adopted as an IEEE standard, the company emerged from stealth in late 2022 with a focus on addressing the escalating challenges in semiconductor design. Traditional SoC architectures are increasingly hitting scalability walls due to rising costs, power consumption, and manufacturing complexities, particularly as AI workloads demand more sophisticated multi-die integrations. Eliyan's innovations aim to bridge these gaps by providing high bandwidth, low latency, and low power interconnect solutions that are compatible with industry standards such as Universal Chiplet Interconnect Express (UCIe) and Open Compute Project (OCP) protocols, while also supporting high bandwidth memory (HBM) applications. The company's core technologies revolve around the NuLink(TM) PHY portfolio and NuGear(TM) chiplet families. The NuLink PHY, for example, has been silicon proven at 64Gbps using a 3nm process, offering up to 2000Gbps/mm of edge bandwidth on standard organic packages, effectively doubling bandwidth and halving power compared to conventional methods. This is achieved without the need for costly advanced packaging like silicon interposers, which reduces manufacturing cycle times, improves yields, and enhances sustainability by cutting waste and energy use. The NuGear family complements this by enabling gearing ratios that optimize die-to-die and chip-to-memory connections, targeting bandwidths from 1.6Tbps to 12.8Tbps. Additionally, Eliyan's Universal Memory Interface (UMI(TM) tackles the "memory wall" in AI chips by providing bi-directional interconnects that boost capacity and efficiency in both standard and advanced substrates. These advancements have been validated through tapeouts on leading processes, including TSMC's 5nm (with first silicon in Q1 2023) and 3nm, underscoring commercial readiness for hyperscalers, AI processors, automotive, and edge applications. Eliyan's funding trajectory reflects growing investor confidence in its potential to disrupt the chiplet ecosystem. The journey began with a $40 million Series A round on November 8, 2022, led by Tracker Capital Management and including Celesta Capital, Intel Capital, Micron Ventures, and other strategic backers. This infusion supported early validation, including a successful 5nm tapeout, and positioned the company for broader commercialization. Key to this round was the emphasis on backward compatibility with UCIe and BoW, enabling practical adoption in sectors facing high costs and yield issues. By March 25, 2024, Eliyan secured a $60 million Series B, co-led by Samsung Catalyst Fund and Tiger Global Management, with participation from Intel Capital, SK Hynix, Cleveland Avenue, Mesh Ventures, and others. This round targeted AI specific challenges, such as memory bandwidth in generative AI, and facilitated a 3nm tapeout for NuLink, projecting massive HBM market growth (331% in 2024 and 124% in 2025 per industry forecasts). An extension or follow-on in August 2024 was noted in some databases, though without a disclosed amount, contributing to a pre latest round total of around $100 million. The most recent milestone is the $50 million strategic Series C, involving a consortium of hyperscalers and AI infrastructure giants: AMD, Arm, Coherent, Meta, Samsung Catalyst Fund, and Intel Capital. This brings Eliyan's total raised to approximately $150 million, with the funds earmarked for accelerating manufacturing, qualification, and ecosystem partnerships. The strategic investor mix, spanning chip designers (AMD, Arm, Intel), optics specialists (Coherent), and end users (Meta), signals broad validation for Eliyan's roadmap in overcoming AI scalability bottlenecks like the memory and I/O walls. Quotes from executives in prior rounds highlight this momentum: Farjadrad has emphasized enabling "the ultimate chiplet systems for the new AI era," while investors like Samsung's Marco Chisari noted the shift toward chiplet architectures driven by generative AI demands. In the broader market context, Eliyan operates in a rapidly evolving landscape where chiplets are seen as essential for sustaining Moore's Law equivalents. The chiplet sector is estimated at $50 billion, with HBM applications adding an $8 billion segment growing at a 50% CAGR. Eliyan's differentiators, such as 4x performance gains and power reductions, position it favorably against reliance on advanced packaging from foundries like TSMC (CoWoS) or Intel (EMIB). However, the field is competitive, with UCIe 2.0 standards driven by collaborators like ASE, Alibaba, Google Cloud, NVIDIA, and Qualcomm potentially commoditizing interconnects. | Funding Round | Date | Amount | Round Type | Lead/Key Investors | Purpose/Key Outcomes | | Series A | November 8, 2022 | $40M | Series A | Tracker Capital (lead), Celesta Capital, Intel Capital, Micron Ventures | Validation and tapeout on 5nm; commercial readiness for UCIe-compatible tech; board addition of Dr. Shaygan Kheradpir | | Series B | March 25, 2024 | $60M | Series B | Samsung Catalyst Fund and Tiger Global (co-leads), Intel Capital, SK Hynix, Cleveland Avenue, Mesh Ventures | Advance AI multi-die architectures; 3nm tapeout for NuLink; address memory wall with UMI | | Series C | January 28, 2026 | $50M | Series C (Strategic) | AMD, Arm, Coherent, Meta, Samsung Catalyst Fund, Intel Capital | Commercialize NuLink and NuGear; expand AI infrastructure partnerships; overcome scalability limits in GenAI systems | | Competitors | Focus | Differentiation from Eliyan | | KiwiMoore | Chiplet interconnect and network acceleration | Full-link services vs. Eliyan's PHY-centric approach | | Kandou AI | D2D PHY IP for AI | Similar interconnect specialists; acquisition targets in consolidation wave | | Alphawave Semi | High speed connectivity | Broader semi portfolio; UK-based with acquisition history | | Ayar Labs | Optical interconnects | Photonics focus vs. Eliyan's electrical PHY | | TSMC (CoWoS) | Advanced packaging | Foundry integrated; Eliyan enables standard organic packages for cost savings | | Intel (EMIB) | Embedded bridging | Proprietary vs. Eliyan's standards compliant (UCIe/BoW) | Looking ahead, Eliyan's integration into ecosystems like Arm Total Design (joined November 2025) and OCP contributions (e.g., BoW 2.1 showcased at OCP Summit 2025) could drive adoption. Recognition as an EE Times Silicon 100 startup in 2025 further bolsters its profile. Potential risks include valuation pressures in a consolidating market, where interconnect specialists are acquisition targets, and dependency on foundry partners like TSMC. If successful, Eliyan could enable core count increases, memory expansions, and TCO reductions, fueling the AI boom while navigating debates over welfare education in tech talent pipelines and energy consumption in data centers.

Eliyan Corporation, a connectivity solutions provider for AI and high-performance computing systems, has secured $50 million in strategic investments from AMD, Arm, Coherent and Meta. Existing investors Samsung Catalyst Fund and Intel Capital also participated in the round. The Santa Clara-based company develops interconnect technology to address scalability and power efficiency challenges in next-generation AI systems. Its NuLink PHY portfolio includes silicon-proven die-to-die connectivity at 64Gbps and chip-to-chip solutions ranging from 32Gbps to 448Gbps. The company's NuGear chiplet families target bandwidth from 1.6Tbps to 12.8Tbps for AI accelerator applications. The funding will accelerate manufacturing and qualification of Eliyan's interconnect intellectual property and chiplet products, expand ecosystem partnerships, and support deployment across AI infrastructure and high-performance computing applications.