Digital Plumber Plumbing the information age
RSS

Vendor Radar

Broadcom

Maker of the Tomahawk and Jericho switch chips used in most data center networks, and owner of VMware.

19stories in the archive
4 up from 1last 7 days · 1 the 7 before
5last 30 days
Jun 13first covered · 2026

Recent activity

Stories per week

Latest stories

10 most recent
  1. Eremons Notes · Oct 6, 2026 · Analysis

    Ethernet gains ground on InfiniBand across AI fabric tiers

    Broadcom Tomahawk 6 and Ultra chips now target scale-up and scale-out workloads, encroaching on NVIDIA's proprietary fabric turf. Data points to stronger Ethernet adoption across all three scaling dimensions, suggesting lower InfiniBand penetration and shift toward multi-vendor Ethernet ecosystems.

    Why it matters Tomahawk 6 Davisson CPO now competes directly with InfiniBand NDR for tightly-coupled training—Ethernet is no longer just a scale-out choice.

    Oct 6, 2026 edition

  2. Yahoo Finance · Oct 2, 2026 · Industry news

    HPE lifts AI networking order target to $3 billion

    HPE raises fiscal 2026 Networks for AI order target to $3 billion. Post-Juniper integration shows 117% networking revenue growth YoY through July 2026; networking now 26% of HPE segment revenue. QFX5250 switches on Broadcom Tomahawk 6 (102.4 Tbps) and MX301 routers shipping to cloud and hyperscaler customers. HPE targeting $800M cost synergies by FY28.

    Why it matters Demonstrates integrated routing/switching/telemetry stacks command premium pricing in AI infrastructure. Distributed AI workloads and data center interconnect drive routing growth at 20%+ CAGR, shifting network capex composition toward layer 3 intelligence and multi-vendor scale-out fabrics.

    Oct 3, 2026 edition

  3. Semiconductor Insight · Oct 2, 2026 · Analysis

    102.4T switch silicon ships as 1.6T Ethernet nears

    Market analysis of 2026 Ethernet switch ASIC landscape covering 800G production deployments, 102.4T silicon (Broadcom Tomahawk 6, Cisco Silicon One G300), and 1.6T-ready platforms. Details 224G SerDes as foundational technology for 1.6T and beyond, with Ethernet Alliance roadmap context.

    Why it matters Synthesizes ASIC and optics roadmaps for 2026-2027 fabric planning, confirming 800G as deployment standard and providing visibility into 1.6T timeline and vendor silicon options.

    Oct 2, 2026 edition

  4. IPinfusion · Sep 24, 2026 · Industry news

    Broadcom Tomahawk 6 reaches volume production at 102.4 Tbps

    Broadcom announced production volume for Tomahawk 6 family switch silicon with 102.4 Tbps capacity and support for 1.6T Ethernet ports. The Tomahawk 6 includes conventional switch variants (BCM78910 and BCM78914) and a co-packaged optics variant (BCM78919 Davisson) with 16 x 6.4 Tbps optical engines.

    Why it matters First-generation 1.6T switch silicon in volume production accelerates hyperscaler adoption of ultra-dense AI fabrics and drives CPO architecture maturity for scale-out deployments.

    Oct 1, 2026 edition

  5. Hengtong · Sep 26, 2026 · Industry news

    Multi-vendor 1.6T Ethernet interoperability validated at OFC 2026

    Multi-vendor 1.6T Ethernet interoperability was validated live at OFC 2026 with 40+ companies demonstrating 200G/lane transceivers and silicon. Keysight and Broadcom showed UEC Link Layer Retry and Credit-Based Flow Control at full 800GbE line rate for the first time publicly.

    Why it matters 1.6T multi-vendor sourcing now carries significantly reduced integration risk; UEC link-layer features enable lossless credit-based forwarding to reduce AI fabric tail latency at scale.

    Sep 29, 2026 edition

  6. Broadcom / VMware Newsroom · Aug 31, 2026

    Broadcom Launches VMware AI Factory: Bare-Metal to Model in Hours via MetalSoft Integration

    Broadcom unveiled VMware AI Factory, a software-defined foundation designed to help enterprises deploy, govern, and secure AI workloads from bare-metal infrastructure through to live model inference. A partnership with MetalSoft delivers integrated bare-metal automation that drops provisioning time from weeks to minutes, unifying software and hardware lifecycle into a single operational model through the VCF management console.

    Sep 3, 2026 edition

  7. Data Center Knowledge · Aug 20, 2026 · Industry news

    AI Data Center Networking: Scale Up, Out, and Across with 102.4T Ethernet

    Vendors including Cisco, Broadcom, and Nvidia are converging on 51.2-102.4 Tbps Ethernet switching silicon with deeper buffering and load-balancing strategies to keep GPU utilization high. Cisco's Silicon One G300 is expected to ship by year-end, marking a critical inflection point where the network becomes a primary determinant of cluster efficiency rather than an afterthought bolted onto compute infrastructure.

    Aug 20, 2026 edition

  8. Yahoo Finance / Arista Networks · Aug 16, 2026 · Vendor release

    Arista Networks Q2 2026: $3.04B Revenue on AI Networking Acceleration, Q3 Guidance Raised to $3.30B

    Arista Networks reported Q2 2026 revenue of $3.04 billion with net income of $1.21 billion, raising Q3 guidance to $3.30 billion driven by accelerating demand for AI-driven network deployments. The company's 7060XE7 series 1.6T switches are set to ship in Q4 2026, competing directly with Broadcom Tomahawk 6-based offerings and DriveNets scheduled Ethernet fabric.

    Aug 16, 2026 edition

  9. Anthropic Blog · Aug 11, 2026 · Industry news

    Anthropic Locks In $71B in Compute Commitments Across AWS, Google, Broadcom Amid Revenue Acceleration

    Anthropic revealed $71 billion in compute commitments across AWS, Google Cloud, and Broadcom partnerships. The company's run-rate revenue surpassed $30 billion, with over 1,000 business customers each spending $1M+ annually on Claude services, demonstrating enterprise demand acceleration.

    Aug 11, 2026 edition

  10. Next Gen Infrastructure (NGI) Research Brief · Aug 4, 2026 · Research

    2026 Data Center Networking Report: NVLink Scales Rack-Wide, Ethernet Reaches 100K+ Clusters, Standards Consolidate

    Comprehensive technical analysis reveals NVIDIA NVLink 5 advancing to 1.8 TB/s per-GPU bandwidth with NVL72 rack-scale systems, while open alternatives (UALink, Ethernet-based ESUN) move from specification to hardware production in late 2026/2027. Co-packaged optics (CPO) and linear pluggable optics (LPO) architectures are becoming standard for 800G+ deployments, with 1.6 Tbps links entering early production.

    Aug 4, 2026 edition

Coverage history

9 earlier stories

August 2026 1

July 2026 6

June 2026 2