Broadcom Unveils Next-Gen “Tomahawk 6-O” Co-Packaged Optics Switch ASIC Platform for AI Networks, Achieving 30% Power Efficiency Boost Towards 2025 Commercial Deployment

July 5, 2026

Broadcom Investor Relations News USA

Overview

Broadcom has announced “Tomahawk 6-O,” a new Co-Packaged Optics (CPO) switch ASIC platform specifically designed for AI networks. This platform aims to accelerate CPO adoption in AI data centers by significantly enhancing power efficiency by up to 30% and dramatically increasing port density. With a detailed roadmap towards commercial deployment in 2025, Tomahawk 6-O is poised to become a foundational technology for next-generation AI infrastructure.

In Depth

Key Findings

Broadcom has unveiled “Tomahawk 6-O,” its next-generation Co-Packaged Optics (CPO) switch ASIC platform, designed to meet the explosive demands of AI networks. This innovative solution promises to simultaneously boost AI data center performance and sustainability by improving power efficiency by up to 30% compared to traditional pluggable optical modules, while substantially increasing switch port density.

Technical / Clinical Details

The Tomahawk 6-O platform integrates Broadcom’s high-performance switch ASIC with its proprietary silicon photonics-based optical engine within the same package, employing CPO technology. This intimate integration minimizes the electrical trace length, leading to dramatic reductions in signal loss and power consumption.
Specifically, by shortening the electrical interconnect between the ASIC and the optical engine to mere millimeters, the power consumption per bit for data transmission is reduced to the picojoule range. This is projected to result in an overall system power efficiency improvement of 30% compared to conventional solutions.
Furthermore, the CPO design eliminates the need for optical transceivers on the switch’s front panel, allowing for a doubling or more of port density. This is crucial for addressing the ultra-high bandwidth requirements for GPU-to-GPU and intra-rack connections within AI clusters.

Background & Context

The rise of generative AI and large language models has driven an unprecedented surge in data center network bandwidth and power consumption. AI/HPC clusters, in particular, require hundreds of thousands of GPUs to be interconnected at ultra-high speeds, pushing conventional electrical wiring and pluggable optical modules to their limits in terms of power, thermal management, and physical space. CPO technology has emerged as a highly anticipated next-generation interconnect approach designed to fundamentally resolve these challenges.

Strategic Significance & Outlook

Broadcom’s announcement of Tomahawk 6-O provides a clear roadmap for the commercialization of CPO technology, with initial deployments expected in major AI data centers by 2025. The adoption of this platform will alleviate bottlenecks in AI supercomputing and accelerate the development of more massive and powerful AI models. Broadcom’s strong market leadership, combined with this innovative CPO solution, positions it to potentially establish the de facto standard for future AI infrastructure design.

Source: https://investors.broadcom.com/news/press-release/XXXXXX/

Get our weekly technology intelligence — free

Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.

Subscribe Free — Weekly Tech Intelligence

By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.