Core Insights - Broadcom is set to launch a new network chip named Thor Ultra, which can connect hundreds of thousands of data processing chips to support AI computing systems [1] - The Thor Ultra chip aims to enhance Broadcom's competitive position against Nvidia in the AI application-specific data center networking space [1] - The chip is designed to facilitate efficient information transmission within data centers, playing a crucial role in connecting AI systems with other components [1] Group 1: Thor Ultra Chip - Thor Ultra will allow operators to deploy a significantly higher number of chips compared to previous generations, enabling the construction and operation of large language models for AI applications [1] - Broadcom has doubled the bandwidth of the Thor Ultra compared to its predecessor, with rigorous testing and auditing conducted from the initial production stages [1][2] - Ram Velaga, Senior Vice President at Broadcom, emphasized the importance of networking in distributed computing systems, indicating that all companies involved in GPU business will seek a foothold in networking [1] Group 2: Wi-Fi 8 Solution - Broadcom announced the industry's first silicon-based Wi-Fi 8 solution aimed at the broadband wireless edge ecosystem, suitable for home gateways, enterprise access points, and smart mobile devices [2] - The Wi-Fi 8 solution is tailored to meet the stringent demands of AI-era edge networks in terms of performance, reliability, intelligence, and energy efficiency [2] - The product lineup includes various chips designed for different applications, such as BCM6718 for home and operator access, BCM43840 and BCM43820 for enterprise access, and BCM43109 for edge wireless terminals like smartphones and laptops [2]
加码AI基础设施生态!博通(AVGO.US)发布新一代网络芯片Thor Ultra 同步首发Wi-Fi 8解决方案