1
Custom channel topology
Bespoke micro-channel fin pitch, depth, and inlet/outlet routing tuned to your specific TDP map and ΔP envelope.
AI Server Liquid Cooling — Phase 1 Primary Solution
AI Server Liquid Cooling for Australian Engineering Teams
Direct-to-chip cold plate kits for NVIDIA GB200, GB300, H200, AMD EPYC SP5, and Intel Birch Stream — the five AI server platforms most deployed in Australian sovereign cloud, HPC refresh (Pawsey, NCI), and edge AI. Cold plate design, engineering, and manufacturing are performed at our IATF 16949-aligned partner facility in Huizhou, Guangdong, China. The Carlton, Victoria team delivers customer engineering support, project coordination, and AS/NZS compliance documentation.
≤ 1×10⁻⁹mbar·L/s helium leak rate sensitivity
2,700 WGB200 Bianca module thermal envelope
2–4 weeksprototype lead time door-to-door to AU
Customprogramme volume agreed at quotation
Process & standards glossary (AU engineering reader):
TLP = Transient Liquid Phase diffusion bonding ·
FSW = Friction Stir Welding ·
CAB = Controlled Atmosphere Brazing ·
PPAP = Production Part Approval Process ·
IATF 16949:2016 (international automotive QMS) ·
ISO 9001:2015 (quality management) ·
AS/NZS 60068-2-14 ≡ IEC 60068-2-14:2009 (thermal cycling endurance) ·
IEC 62619:2017 (industrial battery safety) ·
AS/NZS 5139:2019 (BESS installation) ·
AS/NZS 4777.2:2020 (grid-connected inverter).
Page reviewed by
ToneCooling Australia Customer Engineering Support Team
·
Carlton, VIC
·
Last reviewed: 28 April 2026
·
About the team →
Engineering Reference & Standards
What standards and platform references apply to AI server liquid cooling in Australia?
ToneCooling Australia supplies direct-to-chip liquid cooling cold plates for Australian AI server programmes. Cold plate kits are engineered for the NVIDIA GB200 NVL72 platform, AMD EPYC SP5 (LGA6096), and Intel Birch Stream sockets. Manifold port-pitch follows the Open Compute Project DC-MHS specification; thermal envelope per ASHRAE TC 9.9 W32 / W40 Datacom guideline; AU compliance scope mapped to Standards Australia AS/NZS publications.
Section 1 of 8
What is direct-to-chip liquid cooling?
Direct-to-chip liquid cooling mounts a copper or aluminium cold plate directly on the GPU or CPU die surface and removes heat using single-phase liquid coolant flow rather than air. The cold plate is the primary heat exchanger between the silicon and the rack-level cooling distribution unit (CDU). For AI server platforms — NVIDIA GB200, GB300, H200, AMD EPYC SP5, Intel Birch Stream — direct-to-chip is now the practical thermal envelope for densities above 25 kW per rack.
Air cooling reaches its hard limit between 8–25 kW per rack — beyond this, the volume of air required to remove heat exceeds what a hot-aisle / cold-aisle configuration can deliver. Single-phase direct-to-chip cooling extends the envelope to 80–120 kW per rack and beyond, by exchanging heat directly at the die surface where 90 %+ of the thermal load is generated. This shifts the thermal challenge from air movement to fluid hydraulics: pressure-drop budget across the cold plate, manifold design, quick-disconnect couplings, and the CDU's heat-rejection capacity to the facility loop.
Where ToneCooling Australia fits in the liquid cooling stack
GPU/CPU
package
package
Cold plate
Hose + QD
Manifold
Rack CDU
ToneCooling Australia supplies the middle three layers — cold plates, hose & QD assemblies, manifold-ready cooling modules. Not consumer PC liquid cooling. Not facility CDUs.
Section 2 of 8
Which AI server platforms does ToneCooling Australia support?
ToneCooling Australia supplies cold plate kits for five AI server platforms: NVIDIA GB200 Grace Blackwell (2,700 W per Bianca module), NVIDIA GB300 Grace Blackwell Ultra (next-generation, manifold-ready), NVIDIA H200 (8-GPU baseboard, 8 × 700 W + 2 × 134 W NVSwitch + 2 × 156 W CPU), AMD EPYC SP5 (LGA6096, supporting Genoa, Bergamo, Genoa-X, Turin), and Intel Birch Stream (Granite Rapids, Sierra Forest, 350–500 W reference). Each platform has an established customer base in Australian sovereign cloud, Pawsey-NCI HPC refresh, and edge AI deployments.
NVIDIA · Bianca module
GB200
Grace Blackwell Superchip. Direct-to-chip cold plate for the 2 × Blackwell GPU + 1 × Grace CPU configuration.
2,700 W envelope · NVL72 ready
NVIDIA · Grace Blackwell Ultra
GB300
Next-generation cold plate kit with manifold-ready architecture for the Grace Blackwell Ultra platform refresh.
Engineering preview · Q3 2026
NVIDIA · 8-GPU baseboard
H200
Cold plate kit for the H200 8-GPU SXM baseboard, including NVSwitch and host CPU thermal coverage.
8 × 700 W + 2 × 134 W + 2 × 156 W
AMD · EPYC LGA6096
SP5
Cold plate for AMD EPYC Genoa, Bergamo, Genoa-X, and Turin generations on the LGA6096 socket reference.
400 W reference TDP
Intel · Birch Stream
Birch Stream
Next-generation Xeon platform cold plate covering Granite Rapids and Sierra Forest CPU configurations.
350–500 W reference TDP
Section 3 of 8
What ΔP budget is typical for AI server cold plates?
Typical pressure drop (ΔP) for AI server cold plates ranges 18.5–35 kPa at the manufacturer-recommended flow rate, depending on platform, channel topology, and inlet temperature. NVIDIA H200 reference designs operate around 18.5 kPa at 8.0 L/min on a 50/50 ethylene-glycol-water (EGW) coolant. GB200 vacuum-brazed configurations are typically 25–35 kPa due to the higher heat flux per Bianca module. Pressure drop can be tuned during DFM iteration to align with your CDU's hydraulic budget.
| Platform | Reference flow rate | Typical ΔP | Coolant | Inlet temperature window |
|---|---|---|---|---|
| NVIDIA H200 (8-GPU baseboard) | 8.0 L/min | 18.5 kPa | 50/50 EGW | 20–32 °C (ASHRAE W32) |
| NVIDIA GB200 (Bianca, vacuum-brazed) | 10–12 L/min | 25–35 kPa | 50/50 EGW or PG25 | 20–40 °C (ASHRAE W40) |
| NVIDIA GB300 (manifold-ready) | 10–12 L/min | 28–38 kPa (preview) | 50/50 EGW | 20–40 °C (ASHRAE W40) |
| AMD EPYC SP5 (LGA6096) | 5–8 L/min | 15–25 kPa | 50/50 EGW | 20–32 °C (ASHRAE W32) |
| Intel Birch Stream | 5–8 L/min | 15–22 kPa | 50/50 EGW | 20–32 °C (ASHRAE W32) |
Pressure drop is the primary hydraulic constraint at the cold-plate level. Too high — your CDU pump curve cannot meet the rack's flow demand at the design point. Too low — channel residence time drops, and convective heat-transfer coefficient suffers. ToneCooling's DFM engineering review (performed at our Huizhou facility, coordinated by the Carlton customer engineering support team) runs conjugate-heat-transfer CFD on your actual coolant chemistry, flow rate, and inlet-temperature window before committing to tooling.
Section 4 of 8
How does ToneCooling validate AI server cold plate quality?
Every AI server cold plate undergoes production-line QA at our Huizhou IATF 16949-aligned facility before shipment to Australia. Standard validation pack: high-sensitivity helium leak detection (≤ 1×10⁻⁹ mbar·L/s), ultrasonic C-scan inspection of the bonded interface, pressure and flow performance testing, environmental reliability testing across −40 °C to 150 °C and 20–98 % RH. Lot-traceable inspection certificates accompany every shipment.
Helium leak detection
Every cold plate is tested with a high-sensitivity helium mass-spectrometer leak detector to a sensitivity floor of ≤ 1×10⁻⁹ mbar·L/s. This is two orders of magnitude tighter than the typical industrial pressure-decay leak threshold and is required for hyperscaler-grade direct-to-chip cooling.
Sensitivity ≤ 1×10⁻⁹ mbar·L/sUltrasonic C-scan inspection
Non-destructive inspection of the bonded interface using ultrasonic C-scan imaging. The output is a 2-D image of the bonded plane — uniform colour indicates void-free bonding, dark patches indicate voids or unbonded zones. Performed on every production unit, not a sampled basis.
100 % production unit inspectionPressure and flow performance
Hydraulic ΔP measured against design specification at the rated flow rate, on the actual coolant chemistry your CDU will run. Pressure cycle testing to verify structural integrity at maximum working pressure plus a safety margin. Burst-pressure testing on sample units per lot.
Working pressure 0.35 MPa · burst 4 MPa typ.Environmental reliability
Thermal cycling endurance per IEC 60068-2-14 / AS/NZS 60068-2-14 across −40 °C to 150 °C, with humidity exposure 20–98 % RH per IEC 60068-2-78. Salt-spray exposure per ISO 9227 / AS 2331.3.1 for coastal Australian deployments.
−40 °C to 150 °C · 20–98 % RHSection 5 of 8
What manufacturing process is used for GB200 and H200 cold plates?
Our flagship process is proprietary TLP (Transient Liquid Phase) diffusion bonding — a 5-minute cycle per cold plate, bond strength 0.85 (vs 0.75 for vacuum brazing), production yield 99 % (vs 90 %). TLP forms a continuous metallurgical grain structure across the bonding interface, eliminating the brittle braze-filler interface layer found in conventional vacuum brazing. For projects requiring it, we also supply vacuum-brazed copper (legacy hyperscaler grade) and friction stir welded (FSW) variants.
| Metric | TLP diffusion bonding (ToneCooling) | Vacuum brazing (industry standard) | Why it matters |
|---|---|---|---|
| Bonding cycle time | ~ 5 minutes / part | ~ 1 hour / part | 12× faster cycle enables higher throughput on the same equipment footprint |
| Bond strength (relative) | 0.85 | 0.75 | Higher bond strength survives more thermal cycles in field deployment |
| Production yield | ~ 99 % | ~ 90 % | 9 percentage-point yield delta translates directly to landed cost |
| Microstructure | Continuous metallurgical grain | Brittle filler-metal interlayer | Eliminates the braze-filler weak point under thermal cycling fatigue |
| Bond verification | 100 % ultrasonic C-scan | 100 % ultrasonic C-scan | Same NDI method; both processes verified non-destructively |
Section 6 of 8
How is AI server cooling different from BESS cooling?
AI server cooling is high-heat-flux per area (700–2,700 W concentrated on a small die), narrow ΔT tolerance (≤ 10 °C across the cold plate), tight pressure-drop budget (CDU-constrained), and uses copper or aluminium cold plates with micro-channel topology. BESS cooling is lower flux, distributed across many cells (15 W per cell across 104 cells), broader ΔT tolerance (cell-to-cell uniformity ≤ 6 °C), looser pressure budget, and uses stamped-and-brazed aluminium plates. Both validate to helium leak ≤ 1×10⁻⁹ mbar·L/s.
| Dimension | AI server cold plate | BESS cold plate (TC-1P104S reference) |
|---|---|---|
| Heat source | Single die / package, 700–2,700 W | 104 distributed cells, ~ 15 W each (1.56 kW total) |
| Heat flux density | High (concentrated) | Low (distributed) |
| Channel topology | Micro-channel fins (< 1 mm) | Stamped channels (3–8 mm) |
| Material | Copper (vacuum braze / TLP) or aluminium | Aluminium AL3003MOD + AL4045 brazing |
| Joining process | TLP diffusion bonding · vacuum brazing · FSW | Stamped-and-brazed (CAB) + laser welding |
| ΔT tolerance | ≤ 10 °C across plate | ≤ 6 °C cell-to-cell uniformity |
| ΔP budget | 15–35 kPa (CDU-constrained) | 10–25 kPa (BMS pump-constrained) |
| AS/NZS context | Data centre installation standards (AS/NZS 3000 facility loop) | BESS installation AS/NZS 5139:2019 + grid-connected inverter AS/NZS 4777 |
Both product families share ToneCooling's core validation methodology — helium leak ≤ 1×10⁻⁹ mbar·L/s, ultrasonic C-scan inspection, thermal cycling per AS/NZS 60068-2-14 — but the channel topology, joining process, and material selection differ fundamentally because the underlying thermal physics differ.
Section 7 of 8
What is the lead time for AI server cold plate prototypes to Australia?
Prototype lead time for AI server cold plates is as agreed at quotation door-to-door from Huizhou to Australian capital cities, Low-MOQ. The cycle is: Carlton-coordinated response on receipt of a complete RFQ (TDP map, target ΔP, flow rate, qualification scope) with technical review by the Huizhou design and engineering team; NDA executed promptly; DFM review session scheduled promptly; tooled prototype manufacturing and shipping within the agreed lead-time envelope. Production scale agreed during quotation with Production lead times agreed at quotation.
| Stage | Lead time | Owner | Deliverable |
|---|---|---|---|
| 1. RFQ acknowledgement | prompt | Carlton VIC customer engineering support team | Receipt confirmation, completeness check on TDP map / ΔP / flow rate / qualification scope |
| 2. Engineering response | prompt | Carlton-coordinated; technical review by Huizhou design and engineering team | Thermal sensitivity table, process recommendation (TLP / vacuum braze / FSW), draft quote in AUD |
| 3. NDA execution | within standard turnaround | Carlton VIC team | Mutual NDA signed; design-phase IP exchange enabled |
| 4. DFM review session | promptly | Carlton + Huizhou joint Zoom session | CFD output, channel-topology proposal, manifold-port alignment, hydraulic budget review |
| 5. Tooled prototype delivery | Agreed at quotation | Huizhou manufacturing → AU customer site | 5 × prototype cold plates with helium-leak certificate, ultrasonic C-scan report, ChAFTA Certificate of Origin |
| 6. Production run | Agreed at quotation | Huizhou manufacturing | 50+ piece production lot, lot-traceable QC certificates per piece |
Section 8 of 8
Can custom AI server cold plates be designed?
Yes. ToneCooling Australia accepts custom AI server cold plate programmes from Australian OEMs and integrators. Cold plate channel-topology design, CFD optimisation, and DFM iteration are performed at our IATF 16949-aligned partner facility in Huizhou, Guangdong, China, coordinated through the Carlton, Victoria customer engineering support team. Custom envelopes accepted: bespoke socket footprints, non-standard manifold orientations, OCP DC-MHS-aligned pitches, and project-specific ΔP / flow rate targets. Low-MOQ for tooled prototype, lead time agreed at quotation.
2
OCP DC-MHS alignment
Manifold-port pitch and quick-disconnect coupling alignment per the Open Compute Project DC-MHS specification.
3
OEM logo / branding
Laser-etched or printed branding on the cold plate body for OEM-finished product programmes.
4
Non-standard sockets
Custom AI accelerator sockets (Cerebras, SambaNova, AMD Instinct MI300X, Tenstorrent) on request — engineering preview.
5
Australian compliance pack
AS/NZS 60068-2-14 thermal cycling reports, ISO 9227 salt-spray for coastal racks, RoHS 2.0 + PFAS-free SGS certification.
6
Volume scaling
Programme volumes from 5 prototype to 50+ production to 5,000+ annual capacity per programme. PPAP Levels 1–5 supported.
Frequently Asked
Questions Australian engineers ask about AI server cooling
What is direct-to-chip liquid cooling for AI servers?
Direct-to-chip liquid cooling mounts a copper or aluminium cold plate directly on the GPU or CPU die surface and removes heat using single-phase liquid coolant flow rather than air. The cold plate is the primary heat exchanger between the silicon and the rack-level cooling distribution unit (CDU). For AI server platforms — NVIDIA GB200, GB300, H200, AMD EPYC SP5, Intel Birch Stream — direct-to-chip is now the practical thermal envelope for densities above 25 kW per rack.
Which AI server platforms does ToneCooling Australia support?
ToneCooling Australia supplies cold plate kits for five AI server platforms most deployed in Australian sovereign cloud, HPC refresh (Pawsey, NCI), and edge AI deployments: NVIDIA GB200 Grace Blackwell (1,200 W + 1,200 W + 300 W per Bianca module = 2,700 W total), NVIDIA GB300 Grace Blackwell Ultra (next-generation, manifold-ready architecture), NVIDIA H200 (8-GPU baseboard, 8 × 700 W plus 2 × 134 W NVSwitch and 2 × 156 W CPU), AMD EPYC SP5 (LGA6096, supporting Genoa, Bergamo, Genoa-X, Turin), and Intel Birch Stream (Granite Rapids, Sierra Forest).
Where are AI server cold plates designed and manufactured?
Cold plate design, engineering, and manufacturing are performed at our IATF 16949-aligned partner facility in Huizhou, Guangdong, China — including channel-topology design, conjugate-heat-transfer CFD simulation, process engineering (TLP diffusion bonding, vacuum brazing, friction stir welding), DFM iteration, and quality validation. The Australian team in Carlton, Victoria delivers customer engineering support, RFQ intake, project coordination, AS/NZS compliance documentation, and pre-sales engineering consultation in Australian Eastern time.
What ΔP budget is typical for AI server cold plates?
Typical pressure drop (ΔP) for AI server cold plates ranges 18.5–35 kPa at the manufacturer-recommended flow rate, depending on platform, channel topology, and inlet temperature. NVIDIA H200 reference designs operate around 18.5 kPa at 8.0 L/min on a 50/50 ethylene-glycol-water (EGW) coolant. GB200 vacuum-brazed configurations are typically 25–35 kPa due to the higher heat flux per Bianca module. Pressure drop can be tuned during DFM iteration to align with your CDU's hydraulic budget.
How does ToneCooling validate AI server cold plate quality?
Every AI server cold plate undergoes production-line QA at our Huizhou IATF 16949-aligned facility before shipment to Australia. Standard validation pack: high-sensitivity helium leak detection (≤ 1×10⁻⁹ mbar·L/s), ultrasonic C-scan inspection of the bonded interface (non-destructive void detection), pressure and flow performance testing (coolant resistance, pressure drop), and environmental reliability testing (thermal cycling −40 °C to 150 °C, humidity 20–98 % RH). Lot-traceable inspection certificates accompany every shipment.
What manufacturing process is used for GB200 and H200 cold plates?
Our flagship process is proprietary TLP (Transient Liquid Phase) diffusion bonding — a 5-minute cycle per cold plate, bond strength 0.85 (vs 0.75 for vacuum brazing), production yield 99 % (vs 90 %). TLP forms a continuous metallurgical grain structure across the bonding interface, eliminating the brittle braze-filler interface layer found in conventional vacuum brazing. For projects requiring it, we also supply vacuum-brazed copper (legacy hyperscaler grade) and friction stir welded (FSW) variants. Bond integrity is verified on every production unit using ultrasonic C-scan inspection.
How is AI server cooling different from BESS cooling?
AI server cooling is high-heat-flux per area (700–2,700 W concentrated on a small die), narrow ΔT tolerance (≤ 10 °C across the cold plate), tight pressure-drop budget (CDU-constrained), and uses copper or aluminium cold plates with micro-channel topology. BESS cooling is lower flux, distributed across many cells (15 W per cell across 104 cells), broader ΔT tolerance (cell-to-cell uniformity ≤ 6 °C), looser pressure budget, and uses stamped-and-brazed aluminium plates. Both validate to helium leak ≤ 1×10⁻⁹ mbar·L/s.
What is the lead time for AI server cold plate prototypes to Australia?
Prototype lead time for AI server cold plates is as agreed at quotation door-to-door from Huizhou to Australian capital cities, Low-MOQ. The cycle is: Carlton-coordinated response on receipt of a complete RFQ (TDP map, target ΔP, flow rate, qualification scope) with technical review by the Huizhou design and engineering team; NDA executed promptly; DFM review session scheduled promptly; tooled prototype manufacturing and shipping within the agreed lead-time envelope. Production scale agreed during quotation with Production lead times agreed at quotation.
Continue Exploring
Related ToneCooling Australia pages
PRODUCT
NVIDIA GPU Cold Plate Kits
GB200, GB300, H200 cold plate kits for Australian AI server programmes — TLP diffusion-bonded copper.
View product → PRODUCTServer CPU Cold Plate Kits
AMD EPYC SP5 (LGA6096) and Intel Birch Stream cold plates with 350–500 W reference TDP envelope.
View product → CAPABILITYTLP Diffusion Bonding
Core manufacturing IP — 5-min cycle, 99% yield, 0.85 bond strength. Ultrasonic C-scan verified.
View capability → RESOURCEAI Server Cooling Design Guide
12-section engineering reference — channel topology, ΔP budget, manifold integration, CDU compatibility.
Read the guide →Send your AI server cold plate brief
A complete RFQ — TDP map per platform, target ΔP and flow rate, CDU pump curve, qualification scope, lead time — receives a Carlton-coordinated response within 24 working hours, with technical review by the Huizhou design and engineering team, including a thermal sensitivity table, process recommendation (TLP / vacuum braze / FSW), and draft quote in AUD.
Request a Quote Email info@tonecooling.com.au