Who makes NVLink copper spine for AI data centers
The copper cable assemblies and backplane cartridges that physically wire the scale-up fabric together inside a rack, most visible in Nvidia's NVL72 architecture.
NVLink copper spine and backplane cartridges are the physical layer under the scale-up fabric of 3.01, wiring GPU compute trays to NVLink switch trays inside a rack — the taxonomy notes roughly 5,000 copper cables in an NVL72-class rack. That cable count is itself the reason the industry is evaluating co-packaged optics (3.08) as a replacement at the next speed generation: as per-GPU bandwidth rises, all-copper backplanes start running into reach and density limits that pushing more copper doesn't solve. Several connector and cable assembly makers supply this content, generally under reference-design partner arrangements with the rack architect rather than as a fully open market. An investor should watch this layer as a bellwether for the copper-to-optical transition inside the rack: continued growth in cable count per rack favors incumbent copper suppliers, while an early shift to CPO would cap that growth sooner than volume trends alone would suggest.
AI delta: Larger scale-up domains push cable count per rack higher, up to the point where copper's reach and density limits at the highest speeds make an optical alternative necessary.
About 5,000 copper cables per NVL72; the reason 1.6T-era racks push toward CPO.
Per rack: ~5,000 copper cables per NVL72 (taxonomy v0.1, from Nvidia rack materials).
Companies on this part (5)
| Company | Role | Ownership | Ticker | Price | 1d | Confidence | Note |
|---|---|---|---|---|---|---|---|
| Amphenol | manufacturer | public | APH | 74.09 USD | +1.4% | reported | |
| Bizlink | manufacturer | public | 3665.TW | 1875 TWD | -4.3% | reported | |
| Foxconn Interconnect | manufacturer | public | 6088.HK | 3.71 HKD | -0.1% | reported | |
| Luxshare | manufacturer | public | 002475.SZ | 50.42 CNY | +7.0% | reported | |
| TE Connectivity | manufacturer | public | TEL | 208 USD | -0.2% | reported |
Who competes here
The same table read the other way: every company above competes for NVLink copper spine / backplane cartridges sockets. Confidence tags matter — an inferred placement is a lead, not a confirmed rivalry.
Seen in real machines
NVLink copper spine / backplane cartridges is named in 1 of the rack teardowns and supercomputer breakdowns on this site. Each line below is a piece of that machine. The companies shown are the ones named on that line, not necessarily the maker of this part.
- Nvidia GB300 NVL72 · rack teardown
- NVLink copper spine and backplane cartridges (~5,000 cables) — Amphenol, TE Connectivity, Foxconn Interconnect, Luxshare, Bizlink
Nearby parts
- Previous in layer: Copper: DAC / ACC / AEC cables
- Next in layer: Fiber, connectors, structured cabling
- Layer: Layer 3 — Networking & interconnect
From the blog
Latest feed
- Commentary: Apple's M8 Ultra AI server plan could pull Apple Silicon into Nvidia's enterprise infrastructure game — DigiTimes, 2026-09-24Apple is reportedly developing an enterprise AI server using its planned M8 Ultra chip and integrating Nvidia's NVLink Fusion interconnect technology.
- ACES targets AI compute, power with cable ramp-up — DigiTimes, 2026-09-20Increased demand from AI server buildouts and cloud providers adopting in-house ASICs is driving the need for high-speed interconnects and transmission cables.
- AI server exports to US slow; electronics parts jump 1.5 times in August — DigiTimes, 2026-09-13Taiwan reported that electronic parts exports reached a record high of $32.22 billion in August 2026, with integrated circuits and printed circuit boards seeing significant growth.
- d-Matrix Joins the NVIDIA NVLink Fusion Platform — ServeTheHome, 2026-09-12d-Matrix plans to utilize NVIDIA's NVLink Fusion technology for its next-generation Raptor AI accelerators to facilitate scaling.
- The 'Copper Wall': inside the physical bottlenecks threatening to melt AI supercomputing boom — DigiTimes, 2026-09-04The industry faces potential bottlenecks related to copper supply and interconnects, which threaten the scalability of AI supercomputing.
- NVIDIA and Mediatek Ink $3.5B Investment Deal, Accelerate NVLink Fusion Adoption — ServeTheHome, 2026-09-01NVIDIA and MediaTek have entered into a $3.5 billion investment agreement to accelerate the adoption of NVLink Fusion technology for high-speed GPU-to-GPU interconnects.
- Nvidia pours $3.5 billion into MediaTek — company will adopt NVLink Fusion for its custom AI accelerators — Tom's Hardware, 2026-09-01Nvidia has invested $3.5 billion in MediaTek, which will utilize NVLink Fusion technology for its custom AI accelerator designs.
- Nvidia targets HBM race with custom NVHBM, expands NVLink Fusion — DigiTimes, 2026-08-31Nvidia is accelerating its memory technology roadmap by developing custom NVHBM memory and expanding its NVLink Fusion interconnect technology to enhance data-center performance.
Questions about NVLink copper spine / backplane cartridges
- What are NVLink copper spine / backplane cartridges?
- The copper cable assemblies and backplane cartridges that physically wire the scale-up fabric together inside a rack, most visible in Nvidia's NVL72 architecture.
- Which companies make NVLink copper spine for AI data centers?
- This catalog lists 5 companies on NVLink copper spine / backplane cartridges, 5 of them with listed equity: Amphenol, Bizlink, Foxconn Interconnect, Luxshare, TE Connectivity.
- How does AI change demand for NVLink copper spine / backplane cartridges?
- Larger scale-up domains push cable count per rack higher, up to the point where copper's reach and density limits at the highest speeds make an optical alternative necessary.
- How many NVLink copper spine / backplane cartridges does a data center need?
- Per rack: ~5,000 copper cables per NVL72 (taxonomy v0.1, from Nvidia rack materials). Figures come from the taxonomy behind this catalog.
- How reliable are these NVLink copper spine / backplane cartridges supplier placements?
- The 5 placements on this part break down as 5 reported. "disclosed" means the company or a filing says so; "reported" comes from trade press or a reference-design list; "inferred" is a taxonomy lead with no citation yet.
Sources: placements without a source URL come from the taxonomy seed and are tagged accordingly. How this data is built.· Markdown · JSON