Tag Archives: CDM

Cooling AI Data Centers

Posted on June 27, 2025 by | 1 comment

How important are AI data centers? In just months, Elon Musk’s xAI team converted a factory outside Memphis into a cutting-edge, 100,000-GPU center for training the Colossus supercomputer—home to the Grok chatbot.

Initially powered by temporary gas turbines (later replaced by grid power), Colossus installed its first 100,000 chips in only 19 days, drawing praise from NVIDIA CEO Jensen Huang. Today, it operates 200,000 GPUs, with plans to reach 1 million GPUs by the end of 2025. [1]

Figure 1 – Elon Musk’s 1 Million Sq Ft xAI Colossus Supercomputer Facility near Memphis, TN. [1]

There are about 12,000 data centers throughout the world, nearly half of them in the United States. Now, more and more of these are being built or retrofitted for AI-specific workloads. Leaders include Musk’s xAI, Microsoft, Meta, Google, Amazon, OpenAI, and others.

High power is essential for such operations, and like computational electronics of all sizes heat issues need to be resolved.

GenAI

A key driver of data center growth is Generative AI (GenAI)—AI that creates text, images, audio, video, and code using deep learning. Chatbots and large language model ChatGPT are examples of GenAI, along with text-to-image models that generate images from written descriptions.

Managing all this is possible from new generations of processors, mainly GPUs. They all draw on higher levels of power and generate higher amounts of heat.

Figure 2 – Advanced AI Processor, the NVIDIA GH200 Grace Hopper Superchip with Integrated CPU to Increase Speed and Performance. [2,3]

AI data centers prioritize HPC hardware: GPUs, FPGAs, ASICs, and ultra-fast networking. Compared to CPUs (150–200 W), today’s AI GPUs often run >1,000 W.  . To handle massive datasets and complex computations in real-time they need significant power and cooling infrastructure.

Data Center Cooling Basics

Traditional HVAC was sufficient for older CPU-driven data centers. Today’s AI GPUs demand far more cooling, both at the chip level and facility-wide. This has propelled a need for more efficient thermal management systems at both the micro (server board and chip) and macro (server rack and facility) levels. [4]

Figure 3 – The Colossus AI Supercomputer Now Runs 200,000 GPUs. It Operates at 150MW Power, Equivalent to 80,000 Households. [5]

At Colossus, Supermicro 4U servers house NVIDIA Hopper GPUs cooled by:

  • Cold plates
  • Coolant distribution manifolds (1U between each server)
  • Coolant distribution units (CDUs) with redundant pumps at each rack base [6]

Each 4U server is equipped with eight NVIDIA H100 Tensor Core GPUs. Each rack contains eight 4U servers, totaling 64 GPUs per rack.

Between every server is a 1U manifold for liquid cooling. They connect with CDUs, heat-exchanging Coolant Distribution Units at the bottom of each rack that include a redundant pumping system. The choice of coolant is determined by a range of hardware and environmental factors.

Figure 4 – Each Colossus Rack Contains Eight 4U Servers, Totaling 64 GPUs Per Rack. Between Each Server is a 1U Manifold for Liquid Cooling. [7]
Figure 5 – The Base of Each Rack Has a 4U CDU Pumping System with Redundant Liquid Cooling. [7]

Role of Cooling Fans

Fans remain essential for DIMMs, power supplies, controllers, and NICs.

Figure 6 – Rear Door Liquid-Cooled Heat Exchangers. [7]

At Colossus, fans in the servers pull cooler air from the front of the rack, and exhaust the air at the rear of the server. From there, the air is pulled through rear door heat exchangers. The heat exchangers pass warm air through a liquid-cooled, finned heat exchanger/radiator, lowering its temperature before it exits the rack.

Direct-to-Chip Cooling

NVIDIA’s DGX H100 and H200 server systems feature eight GPUs and two CPUs that must run between 5°C and 30°C. An AI data center with a high rack density houses thousands of these systems performing HPC tasks at maximum load. Direct liquid cooling solutions are required.

Figure 7 – An NVIDIA DGX H100/H200 System Featuring Eight GPUs [8]
Figure 8 – The NVIDIA H100 SmartPlate Connects to a Liquid Cooling System to Bring Microconvective Chip-Level Cooling That Outperforms Air Cooling by 82%. [9]

Direct liquid cooling (cold plates contacting the GPU die) is the most effective method—outperforming air cooling by 82%. It is preferred for high-density deployments of the H100 or GH200.

Scalable Cooling Modules

Colossus represents the world’s largest liquid-cooled AI cluster, using NVIDIA + Supermicro technology. For smaller AI data centers, Cooling Distribution Modules (CDMs) provide a compact, self-contained solution.

Figure 9 – The iCDM-X Cooling Distribution Module from ATS Includes Pumps, Heat Exchanger and Liquid Coolant for Managing Heat from AI GPUs and Other Components. [10]

Most AI data centers are smaller, and power and cooling needs are lower, but essential. Many heat issues can be resolved using self-contained Cooling Distribution Modules.

The compact iCDM-X cooling distribution module provides up to 1.6MW of cooling for a wide range of AI GPUs and other chips. The module measures and logs all important liquid cooling parameters. It uses using just 3kW of power, and no external coolant is required.

These modules include:

•         Pumps

•         Heat exchangers

•         Cold plates

•         Digital monitoring (temp, pressure, flow)

Their sole external component is one or more cold plates removing heat from AI chips. ATS provides an industry-leading selection of custom and standard cold plates, including the high-performing ICEcrystal series.

Figure 10 – The ICEcrystal Cold Plates Series from ATS Provide 1.5 kW of Jet Impingement Liquid Cooling Directly onto AI Chip Hotspots.

Cooling Edge AI and Embedded Applications

AI isn’t just for big data centers—edge AI, robotics, and embedded systems (e.g., NVIDIA Jetson Orin, AMD Kria K26) use processors running under 100 W. These are effectively cooled with heat sinks and fan sinks from suppliers like Advanced Thermal Solutions. [11]

Figure 11 – High Performance Heat Sinks for NVIDIA and AMD AI Processors in Embedded and Edge Applications. [11]

NVIDIA also partners with Lenovo, whose 6th-gen Neptune cooling system enables full liquid cooling (fanless) across its ThinkSystem SC777 V4 servers—targeting enterprise deployments with NVIDIA Blackwell + GB200 GPUs. [12]

Figure 12 – Lenovo’s Neptune Direct Water Cooling Removes Heat from Power Supplies, for Completely Fanless Operation. [12]

Benefits gained from the Neptune system include:

  • Full system cooling (GPUs, CPUs, memory, I/O, storage, regulators)
  • Efficient for 10-trillion-parameter models
  • Improved performance, energy efficiency, and reliability

Conclusion

With surging demand, AI data centers are now a major construction focus. Historically, cooling problems are the #2 cause of data center downtime (behind power issues). With the high power needed for AI computing, these builds should carefully fit with their local communities in terms of electrical needs and sources, and water consumption. [13]

AI workloads will increase U.S. data center power demand by 165% by 2030 (Goldman Sachs), with nearly double 2022 levels (IBM/Newmark). Sustainable design and resource-conscious cooling are essential for the next wave of AI infrastructure. [14,15]

References

1. The Guardian, https://www.theguardian.com/technology/2025/apr/24/elon-musk-xai-memphis

2. Fibermall, https://www.fibermall.com/blog/gh200-nvidia.htm

3. NVIDA, https://resources.nvidia.com/en-us-grace-cpu/grace-hopper-superchip?ncid=no-ncid

4. ID Tech Ex, https://www.idtechex.com/en/research-report/thermal-management-for-data-centers-2025-2035-technologies-markets-and-opportunities/1036

5. Data Center Frontier, https://www.datacenterfrontier.com/machine-learning/article/55244139/the-colossus-ai-supercomputer-elon-musks-drive-toward-data-center-ai-technology-domination

6. Supermicro, https://learn-more.supermicro.com/data-center-stories/how-supermicro-built-the-xai-colossus-supercomputer

7. Serve The Home, https://www.servethehome.com/inside-100000-nvidia-gpu-xai-colossus-cluster-supermicro-helped-build-for-elon-musk/2/

8. Naddod, https://www.naddod.com/blog/introduction-to-nvidia-dgx-h100-h200-system

9. Flex, https://flex.com/resources/flex-and-jetcool-partner-to-develop-liquid-cooling-ready-servers-for-ai-and-high-density-workloads

10. Advanced Thermal Solutions, https://www.qats.com/Products/Liquid-Cooling/iCDM

11. Advanced Thermal Solutions, https://www.qats.com/Heat-Sinks/Device-Specific-Freescale

12. Lenovo, https://www.lenovo.com/us/en/servers-storage/neptune/?orgRef=https%253A%252F%252Fwww.google.com%252F

13. Deloitte, https://www2.deloitte.com/us/en/insights/industry/technology/technology-media-and-telecom-predictions/2025/genai-power-consumption-creates-need-for-more-sustainable-data-centers.html

14.GoldmanSachs, https://www.goldmansachs.com/insights/articles/ai-to-drive-165-increase-in-data-center-power-demand-by-2030

15. Newmark, https://www.nmrk.com/insights/market-report/2023-u-s-data-center-market-overview-market-clusters

1 Comment

Posted in Artificial Intelligence, Cabinet, Chassis, Cold Plates, datacenter, Heat Exchangers, Heat Sinks, Jet Impingement, Liquid Cooling, NVIDIA, Thermal Design, thermal management, Uncategorized

Tagged , , , , , , , , , , ,

Cooling Hot Electronics with Cold Plates

Posted on May 14, 2025 by | 2 comments

Cold plates have been used for thermal management since the Apollo moon missions in the 1960s. Today, they serve a critical role in cooling high-performance electronics across many industries.

Power electronic devices generate significant heat, and if their chips exceed safe temperature limits, system reliability and longevity are compromised. Effective thermal management is essential, as lowering a chip’s junction temperature by just 10°C can double its operational life.

Cold plates offer highly efficient, localized cooling by transferring heat from hot components—such as power semiconductors—into a liquid coolant flowing through the plate. The heated liquid then moves to a remote heat exchanger, where it cools before recirculating back to the cold plate.

Compared to forced-air cooling, cold plates deliver superior thermal performance. They are typically smaller, quieter, and lighter than fan-based systems, making them an attractive solution in compact or noise-sensitive environments.

Most cold plates consist of thin-walled aluminum or copper blocks with internal channels or tubing for coolant flow. As liquid moves through the plate, it absorbs heat from the attached components and carries it away for external dissipation. Modern designs often use mini-channels instead of traditional tubing. These intricate internal pathways maximize surface contact with the coolant, improving heat transfer and cooling efficiency.

Figure 1 – Cold Plates are Part of a Liquid Cooling Loop that Includes a Pump for Fluid Circulation and a Heat Exchanger to Remove Heat from the Flowing Coolant. [ATS]

More advanced cold plates feature mini-channels in place of tubing. These designs can better match applications and more efficiently transfer heat into the coolant.

Figure 2 – A Cold Plate with Internal Mini-Channels Provides a High Rate of Thermal Transfer to Remove More Component Heat. [ATS]

Tubed Cold Plates Cool Hot Electronics

Embedded tube designs are the simplest cold plates. They feature a stainless steel or copper tube coiled and set into grooves inside a metal base plate. The tubes can be routed in different pathways to optimize thermal transfer performance. The flowing coolant moves heat from the component, away from the cold plate and over to a heat exchanger where it is cooled before being pumped back to the plate.

Figure 3 – A Tubed Cold Plate Can Consist of Copper or Stainless-Steel Tubing Pressed or Embedded in a Metal Plate. [ATS]

These tubed cold plates are cost-effective solutions for low- to moderate-power applications and are ideal for use in automotive, instrumentation, and UPS systems. ATS offers models ranging from 57–914 mm in length and 57–198 mm in width, with push-to-connect fittings for easy installation.

A variation of this design features thermally conductive epoxy completely covering the tubing and flush with the plate’s surface. This not only improves thermal contact but also provides environmental protection by sealing the tube within the plate.

Figure 4 – A Cold Plate’s Tubing Can be Buried and Covered with a Thermally Conductive Epoxy Layer. [1]

Custom Cold Plates Provide Best Cooling Solutions

For more demanding applications—such as cooling BGAs, LEDs, or high-power modules—custom cold plates offer the best performance. These can include embedded tubing or submerged internal fins, which increase surface area and create turbulence to enhance heat transfer.

Figure 5. Custom Liquid Cold Plate with Inlaid Copper Tubing Provides Heat Transfer Away from Hot Electronics [2]

One example uses tightly spaced aluminum pin fins to generate turbulence with minimal pressure drop, achieving high thermal performance while keeping the plate compact. Another design incorporates internal turbulators tailored for IGBT modules, further improving coolant flow and heat dissipation.

ATS designs and manufactures custom cold plates in collaboration with customers or based on in-house thermal analysis. These designs can include complex internal geometries such as microchannels or gyroid lattices, like the 3D-printed cold plate created for race car IGBT cooling—an approach that improves flow guidance while reducing weight.

Figure 6. Close-spaced Pin Fins with Complex Geometry Create Turbulence with Low Flow Rate Values Inside Submerged Fin Cold Plates. [3]
Figure 7. A Custom IGBT-Cooling Cold Plate Features Internal Turbulators to Optimize Coolant Turbulence and Heat Transfer. [4]

ATS constructs cold plates to customer designs and those developed in partnership with our own thermal engineers. ATS coolant-based cold plates can include internal tubing and microchannels in closed loop systems.

Figure 8. Dual-Sided Cold Plates Cool Components on Both of Their Mounting Sides. [ATS]

Dual-sided high-flow cold plates provide equal cooling performance for components mounted on both sides of the plate, increasing efficiency, space savings and economy. The cold plates can be used with coolant flow rates up to 4 gal/min, and provide thermal resistance as low as 0.0021°C/W. [5]

Figure 9 – This 3D-Printed Metal Cold Plate Cools IGBTs on a Race Car. It Features a Gyroid Lattice That Guides Internal Coolant Flow While Reducing the Overall Weight [6]

DIY Cold Plates Optimize Component Cooling

ATS also offers DIY (do it yourself) cold plates with modular dimensions and pre-defined drill zones. Engineers can customize mounting locations to match specific components. Once the ideal configuration is determined, ATS can mass-produce the cold plate to match exact specifications.

Figure 10 – Do It Yourself Cold Plates from ATS Feature Drill Zones for Precision Matching to Heat Sources, and No Drill Zones to Protect Internal Coolant Flow Channels. [ATS]

The Complete Liquid Cooling Loop

Cold plates are just one part of a complete liquid cooling system. As electronics demand more efficient cooling, liquid-based systems are increasingly replacing air-based solutions. A functional loop includes a pump, reservoir, and heat exchanger to remove heat from the circulating fluid.

Figure 11 – A Liquid Cooling Loop Featuring Cold Plates. This is Implemented in Avionics on F-16 Fighting Falcon Jets. [7]

These systems are becoming more cost-effective and safer, making liquid cooling viable for a broader range of applications. Cold plates serve as a critical stage in these loops, offering simplicity, versatility, and high thermal performance.

Plug and Play Liquid Loops

For streamlined implementation, ATS offers the Industrial Cooling Distribution Module™ (iCDM™)—a fully integrated liquid cooling loop in one compact, portable unit. It includes the pump, heat exchanger, reservoir, precision controls, and monitoring displays, eliminating the need to purchase and configure components separately.

Figure 12 – The New Industrial Cooling Distribution Module Contains the Pump, Heat Exchanger, Reservoir and Controls for Managing Coolants in Liquid Cooling Loops. [ATS]

The iCDM connects directly to cold plates or chassis-based cooling systems. It supports models with cooling capacities of 10 kW and 20 kW, each holding up to 2 liters of coolant. The system is compatible with a wide range of wetted materials, allowing flexible deployment across industries. A next-generation iCDM, available soon, is fully automated, with significantly increased cooling capacities up to 1.4MW.

Conclusion

Cold plates can provide essential electronics cooling because of their design versatility and the power of liquid cooling. AI chip cooling cold plates will soon be available for this growing and demanding arena.

ATS engineers are experts in matching thermal solutions to system needs, offering a wide portfolio that includes cold plates, vapor chambers, coolant chillers, and complete liquid loop systems. Whether liquid cooling is the best solution depends on the specific application, and ATS provides detailed analysis to help customers make informed decisions.

References

  1. Wakefield Thermal, http://www.wakefield-vette.com/products/liquid-cooling/liquid-cold-plates/standard-liquid-cold-plates.aspx
  2. Baknor, https://www.baknorthermal.com/liquid-cold-plates-various-channel-options/
  3. COOLTECH, http://www.cooltech.it/products/liquid-cold-plates/
  4. Boyd Corp., https://www.boydcorp.com/thermal/liquid-cooling-systems/liquid-cold-plates.html
  5. ATS, https://www.qats.com/Products/Liquid-Cooling/Dual-Sided-Cold-Plates
  6. nTop, https://www.ntop.com/resources/case-studies/cold-plate-automotive-power-electronics/
  7. ThermOmegaTech, https://www.tot-ad.com/avionics-cooling/

2 Comments

Posted in Cold Plates, Liquid Cooling

Tagged , , , , , , , , , , ,