Cisco Systems has introduced its latest innovations, the Silicon One P200 networking chip and the Cisco 8223 routing system, to meet the surging demands of artificial intelligence (AI) workloads. Designed to enhance speed, efficiency, and scalability, these technologies promise to redefine how data centers connect and share massive volumes of AI-driven data across continents. Hyperscale cloud providers, including Microsoft Azure and Alibaba Cloud, have already begun adopting the new systems.
The Silicon One P200 represents a major engineering breakthrough. By replacing 92 separate chips with a single, deep-buffer routing silicon, it significantly reduces hardware complexity while delivering over 65% energy savings compared to traditional routers—a crucial advantage as hyperscalers face mounting power and cooling challenges. The chip supports an interconnect bandwidth exceeding 3 exabits per second, ensuring AI models can operate seamlessly across globally distributed data centers without compromising performance.
Complementing the chip is the Cisco 8223 router, purpose-built for AI-scale computing. Capable of processing more than 20 billion packets per second, the 8223 supports 64 ports of 800G, making it the highest-density fixed router available today. Its deep buffering design helps manage traffic surges during intensive AI model training, maintaining system stability and speed. Cisco also emphasizes energy efficiency, claiming the 8223 is the most power-efficient 3RU router in its category, offering switch-like performance with minimal energy use.
Security and adaptability are central to the new system’s design. The 8223 integrates line-rate encryption and post-quantum resilient algorithms, alongside real-time performance monitoring to safeguard sensitive AI data as it moves between facilities. “AI compute is outgrowing the capacity of even the largest data center,” said Martin Lund, Executive Vice President of Cisco’s Common Hardware Group. “With the Cisco 8223 powered by the Silicon One P200, we’re enabling the secure, large-scale connectivity needed for distributed AI architectures.”
Industry leaders have praised the development. Microsoft’s Dave Maltz highlighted how faster networks with more buffering are vital for expanding AI and cloud operations, while Alibaba Cloud’s Dennis Cai called the P200 “a breakthrough in high bandwidth and power efficiency.” The products, launching first for open-source SONiC environments with IOS XR support coming soon, underline Cisco’s growing strength in the AI networking chip market—a space where it’s competing with the likes of Broadcom. Experts believe these innovations mark a new era where infrastructure, not just software, will determine how far AI technology can scale.
source: techeconomy
