Key Takeaways:
- Broadcom has introduced new interconnect and packaging IP designed to support massive AI workloads, including up to 200 Gbps per lane optical links.
- Its Tomahawk 6 switches and Co-Packaged Optics (CPO) target hyperscale inference clusters with improved bandwidth and energy efficiency.
- A new 3.5D packaging blueprint, XDSiP, enables tight chiplet integration across compute and I/O dies for high-performance AI accelerators.
- Broadcom is building a roadmap toward 400 Gbps CPO technology by 2028, signaling a long-term play in advanced AI system design.
- These developments position Broadcom as a key player in AI infrastructure, with an integrated approach challenging NVIDIA’s standalone GPU-stack model.
Broadcom is moving aggressively to define the next generation of AI infrastructure with high-bandwidth optical interconnects and advanced packaging technologies designed for large-scale compute environments. The company has launched a new suite of intellectual property aimed at enabling data center operators and semiconductor designers to build powerful, tightly integrated AI systems.
Central to this offering is its latest switch platform, Tomahawk 6, equipped with Co-Packaged Optics (CPO). This technology allows optical transceivers to be embedded directly with the switch ASIC, removing the need for traditional pluggable modules. With each optical lane supporting speeds of up to 200 Gbps, Broadcom says its architecture can support dense AI clusters with over 500 GPUs communicating at terabyte-scale bandwidths.
Broadcom is also introducing its 3.5D packaging model, called XDSiP. This enables chipmakers to stack compute and I/O chiplets face-to-face using hybrid copper bonding. The goal is to minimize interconnect delay and power consumption while maximizing performance and flexibility. By licensing this IP, Broadcom allows customers to mix process nodes and create application-specific designs optimized for AI training and inference.
What sets Broadcom apart is its integrated approach. Unlike traditional models that separate compute and network infrastructure, Broadcom is building toward a unified stack—combining connectivity, switching, and packaging. This not only reduces bottlenecks but allows hyperscale operators to better tune systems for AI workload demands.
Looking ahead, Broadcom plans to scale its CPO technology further, targeting 400 Gbps per lane capability by 2028. If realized, that would enable next-generation optical fabrics capable of supporting emerging AI and HPC architectures with far greater efficiency.
These moves signal Broadcom’s intent to compete more directly in the AI infrastructure market—an arena long dominated by GPU vendors. But where companies like NVIDIA lead with monolithic acceleration hardware, Broadcom is offering modularity, speed, and interoperability as differentiators.
For hyperscalers and chip designers seeking more flexibility, lower latency, and high-volume bandwidth, Broadcom’s platform may serve as a compelling alternative. With a growing share of inference and edge workloads moving to customized silicon, the timing of this push could prove strategic.
Learn how AI Agents can supercharge your company’s profits and productivity at TMC’s AI Agent Event, Sept 29-30, 2025 in DC.
If you liked this post, you’ll love one of the the leading global business communications and technology events since 1999, the ITEXPO #TECHSUPERSHOW, Feb 10-12, 2026 Fort Lauderdale, Florida.
Don’t forget the collocated MSP Expo – just for managed service providers!
Aside from his role as CEO of TMC and chairman of ITEXPO #TECHSUPERSHOW Feb 10-12, 2026, Rich Tehrani is CEO of RT Advisors and a Registered Representative (investment banker) with and offering securities through Four Points Capital Partners LLC (Four Points) (Member FINRA/SIPC). He handles capital/debt raises as well as M&A. RT Advisors is not owned by Four Points.
The above is not an endorsement or recommendation to buy/sell any security or sector mentioned. No companies mentioned above are current or past clients of RT Advisors.
The views and opinions expressed above are those of the participants. While believed to be reliable, the information has not been independently verified for accuracy. Any broad, general statements made herein are provided for context only and should not be construed as exhaustive or universally applicable.
Portions of this article may have been developed with the assistance of artificial intelligence, which may have contributed to ideation, content generation, factual review, or editing
