Broadcom Unveils 200G/Lane Co-Packaged Optics

On May 15, Broadcom announced the launch of its third-generation single-lane 200G Co-Packaged Optics (CPO) product line, marking a major milestone in CPO technology development. This latest innovation not only delivers a breakthrough in 200G-per-lane optical interconnects but also showcases the growing maturity of Broadcom's second-generation 100G/lane CPO platform. Key advancements were highlighted in areas such as OSAT manufacturing, thermal design, fiber routing, processing techniques, and overall production yield.

With increasing adoption across the technology ecosystem, Broadcom's CPO platform continues to gain momentum, playing a pivotal role in enabling scalable horizontal and vertical expansion of AI infrastructure.

Broadcom established its leadership in CPO technology in 2021 with the debut of the first-generation Tomahawk 4-Humboldt chipset. This pioneering solution enabled early development across the CPO ecosystem, positioning Broadcom ahead of the curve. The platform introduced several critical innovations, including high-density optical engine integration, edge coupling, and detachable fiber connectors.

Building on this foundation, the second-generation Tomahawk 5-Bailly (TH5-Bailly) chipset became the industry's first mass-produced CPO solution. Broadcom focused on scalable manufacturing and automated testing to support future product generations. Through the deployment of its 100G/lane CPO product line, Broadcom gained deep expertise in CPO system architecture—achieving seamless integration of optical and electrical components while maintaining industry-leading low power consumption.

With the launch of its third-generation 200G/lane CPO solution and ongoing development of fourth-generation 400G/lane technology, Broadcom is setting new standards for high-bandwidth, energy-efficient optical interconnects.

The 200G/lane platform is engineered for next-generation high-radix networks, enabling both vertical and horizontal scaling. These networks require the same level of reliability and power efficiency as traditional copper interconnects, especially to support vertical scaling beyond 512 nodes and accommodate the increasing complexity of large-scale AI models. Broadcom’s third-generation solution is specifically designed to address interconnect scaling challenges such as link jitter and operational disruptions, which can significantly impact cost efficiency.