EN|DE
Request a quote
EN|DE
Request a quote
As rack density rises (AI/GPU, HPC), liquid cooling in the data center is becoming the standard. Netonx fluid connectors – leak-free quick disconnects to the UQD standard, made by an NVIDIA-qualified manufacturer – connect direct-to-chip cooling, CDU lines, rack manifolds and GPU trays safely, serviceably and during operation (hot-plug).
View products & specsData-center liquid cooling uses UQD (Universal Quick Disconnect) fluid connectors to move coolant leak-free between cold plates, CDU lines, rack manifolds and GPU trays. Netonx supplies UQD/UQDB, blind-mate and MQD quick disconnects from NVIDIA-qualified manufacturing — leak-free (non-spill / dry-break), >5,000 mating cycles — as a European second source to Stäubli, Parker and CPC. An NVIDIA GB200 NVL72 rack needs roughly ~126 UQD coupling pairs; GB300 rises to ~250 pairs.
What is a UQD (Universal Quick Disconnect)? A UQD is a standardized, leak-free quick-disconnect fluid coupling defined under the Open Compute Project (OCP). It connects and separates coolant lines in direct-to-chip (D2C) cooling quickly and drip-free, even under system pressure, so servers and trays can be serviced without draining the loop.
Leak-free quick disconnects for cold plates and server loops.
Distributor and CDU connection with high flow capacity and tightness.
Blind-mate couplings for hot-swap without manual alignment.
UQD-compliant, >5,000 mating cycles, cleanroom assembly, European delivery.
UQD demand grows with every GPU generation. In the GB200 class an NVL72 rack uses on the order of ~126 coupling pairs (18 compute trays + 9 switch trays). With GB300 – an individual cold plate per chip – market reports estimate ~250 pairs per rack. The couplings also get smaller, raising the bar on manufacturing tolerance and sealing.
| Feature | GB200 (NVL72) | GB300 (NVL72) |
|---|---|---|
| Cooling concept | large-area cold plate | individual cold plate per chip |
| Coupling pairs / compute tray | ~6 | ~12–14 |
| Coupling pairs / rack (≈) | ~126 | ~250 |
| UQD form factor | UQD / UQDB (OCP) | more compact UQD generation |
Figures are industry estimates from market reports and serve as orientation; the exact quantity depends on the specific tray/manifold design. Reference designs commonly use OCP-standard UQD from established makers (CPC, Stäubli, Parker). Read the full GB200/GB300 analysis → · Next generation: Vera Rubin (VR200 NVL72) liquid cooling →
UQD demand per rack rises significantly from GB200 to GB300. Anyone supplying liquid-cooling projects needs a qualified second source – compatible with the OCP standard, with proven sealing and sufficient volume. Netonx supplies UQD-compliant quick disconnects from NVIDIA-qualified production as an alternative / second source, with a European point of contact.
Seamless integration into data centers built on NVIDIA reference designs – UQD standard instead of vendor lock-in.
Optimized flow for D2C on HGX baseboards and GPU trays – higher pump efficiency, better PUE.
Non-spill / dry-break; material and seal selection (stainless vs. aluminium, EPDM vs. FKM) matched to the coolant.
Guide: UQD vs. proprietary → · Test report SFC (DN50) → · UQD connectors explained → · UQD per NVL72 rack →
Tell us your rack architecture, cold-plate interface, CDU flow and volume – we will respond with a UQD-compliant configuration and European lead time.
Contact Netonx Europe