<p>➀ Broadcom launched the Tomahawk Ultra, a 51.2Tbps Ethernet switch chip optimized for small packet sizes and ultra-low latency (250ns), targeting HPC and AI clusters;</p><p>➁ The chip challenges NVIDIA's NVLink and InfiniBand by offering 'Scale-Up Ethernet' (SUE) with features like in-network collectives, compressed headers, and lossless fabric technology;</p><p>➂ Pin-compatible with Tomahawk 5 and shipping alongside Tomahawk 6, it emphasizes scalability for AI workloads while maintaining Ethernet compliance.</p>
Related Articles
- This is the Massive NVIDIA 800G OSFP to 2x 400G QSFP112 Passive Splitter DAC Cable9 days ago
 - Trump says Intel has made a 'fortune' and America has made $40 billion after the US invested in ailing chipmaker — Intel has gained $73 billion in market cap since investment14 days ago
 - The Massive Size of Broadcom Tomahawk 6 102.4T Switch Chips and Tomhawk Ultra Bonus15 days ago
 - Frore's new LiquidJet coldplates are equipped to handle the spiralling power demands of future AI GPUs — built to handle up to 4.4Kw TDPs, solution could be deployed in power-hungry Feynman data centers20 days ago
 - CSP capex $420bn this year; $520bn next year21 days ago
 - Broadcom Tomahawk 6 – Davisson 102.4T Switch with Co-Packaged Optics Shipping22 days ago
 - Microsoft deploys world's first 'supercomputer-scale' GB300 NVL72 Azure cluster — 4,608 GB300 GPUs linked together to form a single, unified accelerator capable of 92.1 exaFLOPS of FP4 inference24 days ago
 - U.S. finally grants Nvidia license to ship AI GPUs to UAE: 500,000 Blackwell GPUs coming to the Gulf region26 days ago
 - STH Q3 2025 Letter from the Editor Trifecta Challengesabout 1 month ago
 - UK energy price may stymie US datacentre investmentabout 1 month ago