AMD vs Broadcom
Relationship
AMD's FY2025 10-K names Broadcom among the ASSP vendors it competes against on data-centre networking/AI-NIC silicon.
2 of 9 capabilities — Shares accelerator silicon and interconnect.
Ludbee capability tags · from the product recordsShared product type — Both ship hardware and platform.
Ludbee product recordsLarger scale — Broadcom: $1.7T market cap, against AMD's $785.1B market cap.
Ludbee scale figures · valuation, market cap or revenue estimateNamed in filing — “from ASSP vendors such as Broadcom Corporation, Marvell Technology Group, Ltd., Analog Devices, Texas Instruments Incorporated, NXP Semiconductors N.V., Qualcomm Incorporated and NVIDIA”
sec.gov · checked 2026-09-18Connect AI clusters — Rivals on this job — Link accelerators, memory and storage inside AI servers and across racks — the interconnect and switching layer of a training cluster.
Ludbee needs vocabulary · the scope on the sourced edgeAligned comparison
Capability overlap
Shared · 2
Not verified for Broadcom · 7
Recorded for AMD. Broadcom’s product records say nothing either way — a missing record is not a missing capability.
Not verified for AMD · 2
Recorded for Broadcom. AMD’s product records say nothing either way — a missing record is not a missing capability.
Products, side by side
Hand-checked pairing
AMD
Platform
An open-source reference stack for running enterprise AI workloads at scale on AMD compute: open-source AI frameworks and generative models connected to an enterprise-ready Kubernetes platform, with AMD Inference Microservices — prebuilt inference containers bundling model, engine and optimised configuration for AMD hardware.
Open-source GPU computing stack of compilers, runtimes and libraries for running AI and HPC workloads on AMD hardware.
A development stack of compiler, runtime and tools for porting pretrained models onto the NPU and integrated GPU of Ryzen AI processors.
An open-source AI inference development stack of compiler, optimised NPU IP and runtime for deploying deep-learning models on AMD adaptive SoCs, FPGAs and Alveo cards.
Hardware
A rack-scale AI system combining 72 Instinct MI455X GPUs with EPYC CPUs and Pensando networking for frontier-model training and large-scale inference.
Data-centre GPU accelerators for training and serving models, sold in OEM servers and rack systems.
System-on-module family for deploying edge and physical AI, with production SOMs and partner-validated carrier systems.
AI NIC family (Pollara 400 at 400 Gbps, Vulcano 800 at 800 Gbps Ethernet) for scale-out inter-GPU communication in AI training and inference clusters, built on the Ultra Ethernet Consortium specification.
Data processing unit family (Salina, Giglio) offloading networking, security and storage services from host CPUs in AI data centres, programmable via the P4-based Pensando software stack.
Second-generation Versal AI Edge adaptive SoC family for embedded AI preprocessing, inference and postprocessing.
Broadcom
Platform
Governance layer for AI agents: each agent is given a verifiable identity with an explicit mission, and a protocol-aware gateway authorizes every tool call against policy before it reaches an enterprise system, leaving an audit trail of what each agent did.
Hardware
Tomahawk and Jericho Ethernet switch silicon used to connect accelerators inside AI clusters.
An 800G PCIe Gen6 Ethernet network interface card built for server-to-server networking in large AI and ML GPU clusters.
A 102.4-Tbps Ethernet switch with co-packaged silicon-photonics optics, built to remove optical-I/O bottlenecks in large-scale AI clusters.
Designs and manufactures custom AI accelerators (XPUs) to a cloud operator's own specification, including the 3.5D packaging that carries them.
No counterpart
AMD sells these in a stack layer with no product recorded for Broadcom yet — nothing on the other side to compare them against.
Infrastructure service
An on-demand cloud service giving developers browser and API access to AMD Instinct MI300X GPU instances for AI, ML and HPC workloads.