Lightmatter has introduced the Passage M1000, a revolutionary 3D Photonic Superchip designed to transform AI infrastructure. This cutting-edge technology offers a record-breaking 114 Tbps total optical bandwidth, making it the fastest AI interconnect available. The M1000 is particularly suited for next-generation XPUs and switches, enabling the most demanding AI applications.
The Passage M1000 is a game-changer for AI infrastructure, enabling seamless communication across thousands of GPUs in a single domain. Its pervasive interposer connectivity, powered by an extensive and reconfigurable waveguide network, ensures high-speed, low-latency data transfers. This technology is designed to eliminate critical data bottlenecks, providing unparalleled efficiency and scalability for advanced AI and high-performance computing workloads.
Lightmatter has collaborated with industry leaders like GlobalFoundries (GF) and Amkor to ensure production readiness for customer designs based on the M1000 reference platform. The M1000 utilizes the GF Fotonix™ silicon photonics platform, which integrates photonic components with high-performance CMOS logic into a single die, creating a production-ready design that can scale effectively with AI demands.
With its advanced photonics-driven approach, Lightmatter aims to overcome power limitations and bandwidth constraints faced by existing AI hardware. The Passage M1000, along with the Passage L200, accelerates advances in AI by enabling larger and more capable AI models to be trained faster than ever before. These innovations not only enhance performance but also reduce energy consumption, making AI computing more sustainable.
For more information, visit Lightmatter's official website.