Eliyan Introduced New Optical Chiplet Platform

The company launched its NuLight system to boost bandwidth and reach for large-scale AI infrastructure.

Updated on Oct. 14, 2026 in Semiconductors

Bold flat-color editorial illustration of a geometric semiconductor wafer stack with a glowing optical channel, representing AI data transmission technology. — Eliyan launched its NuLight optical chiplet platform, designed to expand bandwidth and reach for large-scale AI data center infrastructure. AI Illustration. Upload story photo >

Live Poll

Do you believe new hardware innovations will effectively solve the scaling challenges of modern AI infrastructure?

Eliyan released the NuLight 2D microVCSEL interconnect platform, which uses optical networking to overcome copper connectivity constraints in AI data centers. The company also began sampling its new Nexus and Oscar chiplets.

Why it matters

The platform aims to scale AI infrastructure across multiple racks by providing higher bandwidth and reach while maintaining copper-like reliability. It addresses physical limitations inherent in traditional copper connections for high-performance computing.

The NuLight platform uses microVCSEL and photodiode arrays to support modular configurations up to 12.8 Tbps. The Nexus chiplet utilizes TSMC 3nm process technology to convert 212G PAM4 host links into 4 x 53G NRZ links.

The players

Eliyan

A technology company based in Santa Clara that specializes in interconnect solutions for high-performance computing and AI chipsets.

The details

The architecture targets an optical reach of approximately 50 meters and features on-the-fly lane repair to redirect traffic if an active link fails. Future iterations of the Oscar interface chip are expected to connect directly to host ASICs via die-to-die technology.

Eliyan introduced the NuLight platform on October 14, 2026.
The technology was unveiled during the OCP Summit 2026 the week of October 14, 2026.

The Tech Race

The introduction of the NuLight platform marks a significant shift as industry players move from traditional copper cabling to optical interconnects for rack-to-rack data transfer. This evolution positions Eliyan to compete directly with emerging high-speed interconnect standards in the AI hardware market.

This technology aims to improve the efficiency and speed of large-scale AI infrastructure, potentially reducing latency for data-intensive applications. Future adoption of these chiplets may lead to faster cloud processing capabilities for consumers.

The takeaway

The move toward optical interconnects represents a vital step in scaling data center bandwidth as demand for AI computing grows. Readers should anticipate faster and more reliable network connections as these optical standards become integrated into high-end hardware.