Companies/Broadcom
Palo Alto, CA, USA
Broadcom
Tomahawk and Jericho switch silicon, plus custom accelerator ASIC design for hyperscalers. The largest merchant supplier of Ethernet switching into AI fabrics.
Systems this company appears in
- Systems
- 6
- Aggregate Rmax
- 205.7 PFlop/s
- Aggregate power
- 0 kW
- First appearance
- 2024
- Most recent
- 2025-11
- Concentration
- 56%
systems it appears in
where disclosed
largest is MAXIMUS-384
Aggregate Rmax is the sum of the systems this company supplies any part of. It is not revenue and not attributable share: an integrator and a memory vendor in the same machine both count its full FLOPS here. Attribution weighted by role is Phase 4.
Role breakdown
| Role | Systems | Rmax in those systems |
|---|---|---|
| Interconnect | 6 | 205.7 PFlop/s |
Appearances
| System | Supplied as | Shipped under | Rmax | Rank |
|---|---|---|---|---|
| MAXIMUS-384 USA · 2025-11 | Interconnect | Broadcom | 114.5 PFlop/s | #24 June 2026 → |
| AI-03 United Arab Emirates · 2025-06 | Interconnect | Broadcom | 54.12 PFlop/s | #47 June 2026 → |
| Vultr AMD GPU Supercompute Cluster (Chicago) USA · 2024-12 | Interconnect | Broadcom | - | unranked |
| IronMan (Dell) USA · 2024 | Interconnect | Broadcom | 3.11 PFlop/s | #380 June 2025 → |
| SAKURAONE Japan · | Interconnect | Broadcom | 33.95 PFlop/s | #49 June 2025 → |
| ByteDance MegaScale Cluster China · | Interconnect | Broadcom | - | unranked |
Parts supplied
| Part | Category | Generation | Released |
|---|---|---|---|
| Broadcom RoCE Ethernet | Interconnect | Tomahawk/NetXtreme | 2023 |
| Broadcom Tomahawk 5 | Networking | Tomahawk | 2022 |
Compute deals
| Date | Role | Counterparty | Type | MW | Value ($bn) | Source |
|---|---|---|---|---|---|---|
| 2026-04-06 | Seller | Anthropic | Chip supply | 3,500 | - | anthropic.comigorslab.de |
| 2025-12-11 | Seller | Anthropic | Chip supply | - | 21 | sdxcentral.com |
| 2025-10-13 | Seller | OpenAI | Chip supply | 10,000 | - | seekingalpha.com |
Headline values are as announced and are not additive. See all compute deals.
Related news
- d-Matrix Announces Squadron Compute Clusters for AI Inference at Data Center Scale · HPCwired-Matrix, built around its own memory-centric Corsair XPU chip, is now shipping full air-cooled, multi-rack "Squadron" inference clusters (up to 1,024 XPUs) built with Supermicro, Arista and Broadcom, already running at its own Arizona site and now shipping to outside customers -- a new merchant-silicon entrant at the inference-accelerator layer this graph tracks, pitched for heterogeneous GPU-plus-XPU deployments rather than as a drop-in GPU replacement.
- Broadcom Rides Rocketing Trend For Custom AI Accelerators · The Next PlatformBroadcom's reported AI revenue and forward guidance point to custom accelerator programs (Google, Anthropic, OpenAI, Meta) scaling into multi-gigawatt annual deployments, which matters for the graph because it marks a structural shift of hyperscaler capacity planning away from merchant GPUs and toward vendor-specific silicon built through Broadcom's packaging and TSMC manufacturing, with HBM supply as the binding constraint on how fast that shift can go.