Cisco is releasing new Silicon One chips, the G300 and P200, to address the growing demand for high-performance networking in AI data centers. The G300 is designed for scale-out within data centers, offering 102.4 Tbps of bandwidth to ensure efficient GPU utilization. The P200 chip is targeted at scale-across applications, enabling high-throughput connectivity between data centers with deep buffering capabilities. These new chips aim to prevent network bottlenecks and maximize the performance of AI clusters. AI
IMPACT These chips aim to improve the efficiency and performance of AI data centers by addressing network bottlenecks.
RANK_REASON This is a product announcement for networking hardware, not a core AI model release or research.
Read on Data Center Knowledge →
- Broadcom
- Cisco
- Dell'Oro Group
- Nick Kucharewski
- NVIDIA
- Sameh Boujelbene
- Silicon One G300
- Silicon One P200
- Spectrum 6 (1025)
- Tomahawk 6
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
