Photonic computing company Lightmatter has broadened its industry coalition for standardizing silicon photonics-ready AI infrastructure, adding nine new members to what is now a 19-company initiative. The group, operating as an official workstream within the Open Compute Project (OCP), aims to create a shared blueprint for co-packaged optics (CPO) that can scale AI clusters from 72 to more than 1,024 nodes.
The initiative, first announced in March 2026 with ten founding members, has attracted additional participation from major hardware and technology players. The expanded roster now includes Celestica, Corning, Dell Technologies, Flex, Foxconn Interconnect Technology, Hyve Solutions, Keysight, Qualcomm Technologies, and Quanta Cloud Technology, alongside the nine newly joined companies. The coalition's growth underscores the urgency across the industry to address interconnect bottlenecks as AI workloads demand ever-greater bandwidth and energy efficiency.
To support the project, Lightmatter has published a 300-page white paper titled "Architecture Vision: Open Silicon Photonics for AI Systems." The document aims to guide the development of a shared CPO reference architecture compatible with Modular Hardware System (MHS) and Open Rack v3 (ORv3) standards, providing hyperscalers with a common framework for transitioning from copper-based interconnects to photonic solutions. Unlike traditional pluggable optics, where transceivers remain separate from main chips, CPO integrates optical engines and switching ASICs on a single packaged substrate, reducing power consumption and improving signal integrity.
"Interconnect has become as fundamental to AI infrastructure performance as compute," said Nick Harris, founder and CEO of Lightmatter. "The industry's shift to silicon photonics will unlock the next major leap in frontier model capabilities. Aligning 19 companies on a system-level architectural vision was an enormous undertaking, and I'm proud of what this coalition has accomplished. This white paper lays the foundation upon which this growing OCP community will develop a shared MHS-based CPO rack architecture that fits into hyperscalers' build flows. The work begins now."
The reference architecture is designed to remain scalable and interoperable across multiple SerDes generations, switch designs, and XPU cycles, ensuring that investments in photonic infrastructure can be carried forward as technology evolves. By standardizing the system-level approach, the initiative seeks to reduce fragmentation in the emerging CPO ecosystem and accelerate adoption among hyperscale data center operators.
Lightmatter, based in California, specializes in photonics-enabled computing that uses light instead of electrical signals, avoiding the heat and resistance limitations of traditional chip architectures. The company recently launched its Passage L20 interconnect in March 2026, succeeding the Passage M1000 and L200 products introduced in April 2025. The new initiative builds on this momentum, positioning silicon photonics as a critical enabler for next-generation AI infrastructure at scale.
The move reflects a broader industry recognition that interconnect performance is becoming as vital as raw compute capacity in AI systems. As frontier models grow larger and more complex, the ability to move data efficiently between nodes will increasingly determine overall system performance. With a unified architectural vision and a growing coalition of industry leaders, the path toward photonic-ready AI data centers is becoming clearer.