Broadcom Unveils Jericho4 Ethernet Fabric Router for AI Infrastructure

Broadcom has begun shipping the Jericho4 ethernet fabric router, a purpose-built platform for next-gen distributed AI infrastructure. Jericho4 accelerates AI workloads by managing massive network traffic with high-bandwidth memory, and interconnects over one million XPUs across multiple data centers. Broadcom designed Jericho4 to break traditional scaling limits with unmatched bandwidth, security, and lossless performance.

Broadcom has begun shipping the Jericho4 ethernet fabric router, a purpose-built platform designed to support the next generation of distributed AI infrastructure. The Jericho4 router is engineered to interconnect over one million XPUs across multiple data centers, accelerating AI workloads by managing massive network traffic with high-bandwidth memory [1].

Key features of Jericho4 include its ability to deliver unmatched bandwidth, security, and lossless performance, breaking traditional scaling limits. The router supports low-power, high-bandwidth memory (HBM) packet buffering, providing up to 160 times more traffic buffering than on-chip memory, which enables zero-packet-loss performance in heavily congested networks [1].

Broadcom's Jericho4 complements its existing Tomahawk Ultra and Tomahawk 6 chips, offering a comprehensive networking portfolio for high-performance computing (HPC) and AI environments. A single Jericho4 system can scale to 36,000 HyperPorts, each running at 3.2 Tbps, with deep buffering, line-rate MACsec encryption, and RoCE transport over distances greater than 100 kilometers [1].

The Jericho4 router is manufactured on a 3nm process, increasing transistor density to support more ports, integrated memory, and greater power efficiency. This makes it ideal for coordinating distributed AI processing across expansive GPU farms [2]. Integrated HBM facilitates real-time, localized congestion management, removing the need for complex signaling across nodes during high-traffic AI operations. Enhanced on-chip encryption ensures secure inter-data center traffic without compromising performance [2].

Broadcom's influence in the data center interconnect space suggests that Jericho4 could reshape AI-focused data center designs, particularly for hyperscalers such as Amazon Web Services, Microsoft Azure, and Google Cloud. Its ability to support long-haul interconnects of more than 100 km allows compute and storage resources to be separated across city-scale distances, creating more flexible and resilient regional AI zones [1].

The inclusion of HBM marks a shift toward 'smart switches', integrating compute, memory buffering, and inline analytics directly into the network fabric. Jericho4 also advances security by supporting native encrypted switching, paving the way for zero-trust networking at the hardware level. With congestion-resistant, high-throughput interconnects, it simplifies AI scaling by reducing reliance on complex overlay networks or proprietary fabrics [2].

Analysts suggest that Jericho4 could significantly impact AI-focused data center designs, particularly for hyperscalers. The combination of Jericho4 with other Broadcom chips will allow enterprises and hyperscalers to invest in more hyperscale data centers across a wider range of geographical locations [1].

References:
[1] https://www.networkworld.com/article/4034136/broadcom-expands-ai-networking-portfolio-with-jericho4-ethernet-fabric-router.html
[2] https://www.hpcwire.com/off-the-wire/broadcom-ships-jericho4-enabling-distributed-ai-computing-across-data-centers/