By Ridhika Basnet · AIM Media House
Lightmatter, the California based silicon photonics and artificial intelligence computing company, partnered with NVIDIA NVLink Fusion ecosystem on Jun 2, 2026. The partnership aims to boost the distribution of high-performance optical connectivity for AI infrastructure.
The company produces silicon photonics platforms intended to link multiple AI processors and support next-generation data center infrastructure.Through the partnership, the company will deliver Co-Packaged Optics (CPO) and Near-Packaged Optics (NPO) products, in line with NVIDIA’s optical and SerDes technologies.
"By integrating Passage CPO solutions with NVIDIA's NVLink Fusion architecture, we are combining the industry's most advanced AI platform and the world's leading interconnect to unleash generations of leading frontier AI models,” said Nick Harris, Ph.D., founder and CEO of Lightmatter, visualizing the collaboration as the next era of AI infrastructure.
Lightmatter's Passage is an innovative silicon photonics platform that replaces traditional, slow copper wires with laser light to move data between AI computer chips.
By integrating optical engines directly onto or alongside processors using near-packaged and co-packaged optics (NPO and CPO), Passage allows thousands to millions of chips to communicate seamlessly at the speed of light.
According to the company, this optical hardware architecture increases bandwidth density and reduces energy consumption, supporting the scaling requirements of large-scale AI infrastructure and high-performance computing systems.
AI Connected by Light The company said that by adjusting their two-way optical link architecture for NVIDIA’s optical and electrical technology, the company is building a single platform for semi-custom AI manufacturers, reducing fiber and connector requirements by half.
This approach allows custom AI chips from different suppliers to connect with NVIDIA switches seamlessly for high-speed data sharing, the company added.
Continue on AIM Media House