Edgecore Networks is using this week’s OCP Global Summit to show its latest AI Ethernet switching hardware, including a 102.4 Tb/s Tomahawk 6 platform and a liquid-cooled version aimed at dense AI racks.
Edgecore’s AIS1600-64O exposes 64 ports of 1.6T Ethernet and uses Broadcom’s Tomahawk 6 silicon. The platform supports 512 logical ports through breakout, 267 MB of shared packet buffer and AI-focused congestion-control features. Edgecore’s current OCP social post also highlights a liquid-cooled AIS1600-64O-FL2L variant alongside an air-cooled design.
Broadcom product manager Pete Del Vecchio explains the Tomahawk 6 architecture and its 102.4 Tb/s AI-networking target.
Broadcom says Tomahawk 6 can connect up to 512 XPUs in a single-hop scale-up network and can support 100,000-plus XPUs in a two-tier scale-out design at 200 Gb/s per link. The chip also supports 100G/200G PAM4 SerDes and optional co-packaged optics.
The bigger story is how fast Ethernet fabrics are being redesigned around AI clusters. BitcoinVersus.Tech has recently covered Broadcom’s OCP networking roadmap, 12.8 Tb/s optics for AI networks and chip-scale optical circuit switching. Edgecore has not announced pricing or a general-availability date for the liquid-cooled model.
BitcoinVersus.Tech Editor’s Note:
We volunteer daily to ensure the credibility of the information on this platform is Verifiably True. If you would like to support to help further secure the integrity of our research initiatives, please donate using the QR code provided
For best QR-code scanning results, please scan this donation code from a desktop or larger screen. Mobile screens may prevent camera recognition.
BitcoinVersus.Tech provides technical education and analysis for informational purposes only.
Leave a Reply