{"slug": "asus-thermal-lab-tour-2026-testing-ai-servers", "title": "ASUS Thermal Lab Tour 2026 Testing AI Servers", "summary": "ASUS gave a tour of its thermal testing lab in Taiwan, where it tests AI servers in environmental chambers that simulate years of thermal stress in compressed timeframes. The lab includes a walk-in chamber for full-rack liquid-cooled systems and an extreme chamber for temperatures from -40°C to 85°C, crucial for validating AI infrastructure like the NVIDIA HGX B200 8-GPU server.", "body_md": "Last quarter, I visited the ASUS server thermal testing lab in Taiwan. This is where the company does its thermal testing for customers, validation, and even longer-term reliability testing. Inside the lab, there are a number of facilities, each with a different focus area. One of the fun parts of this lab is that it is where ASUS actually bakes its servers to simulate different operational stresses. Let us get to it.\n\nAs a quick note, ASUS sponsored our visit, so we have to say this is sponsored.\n\n## ASUS Walk-in Environmental Chamber\n\nAI server development moves quickly enough that vendors need facilities capable of simulating years of thermal stress in compressed timeframes. Knowing how a server’s components will fare for 5 years is useful, but by then there will be (several) new generations of AI servers. To accelerate this and simulate different customer environments, the industry uses environmental chambers to reduce the time required to collect this data. Also, these types of chambers are used to validate heatink designs, fans, and so forth.\n\nThis walk-in environmental chamber behind me handles standard operating range validation for both individual servers and complete racks. ASUS runs tests from 25 degrees Celsius up to 45 degrees Celsius, covering the typical data center ambient conditions most enterprises will encounter.\n\nASUS’s NVIDIA HGX B200 8-GPU “Blackwell” generation server currently occupies this chamber during our visit. The space is large enough to accommodate full rack configurations rather than just single nodes, which matters for liquid-cooled systems where coolant distribution and manifold behavior change at rack scale. Testing a complete rack reveals thermal interactions between nodes that single-server validation cannot capture. Still, sometimes, you just need to test a single server. One of today’s 8-GPU server nodes can use the power of two older generation 208V/30A racks. In a way, this is like testing two racks worth of 2015 gear in a single machine.\n\nMost vendors test individual air-cooled servers in environmental chambers. ASUS designed this one to accept full racks because liquid cooling introduces dependencies between nodes that do not exist in only air-cooled deployments. Coolant temperature, flow rate, and pressure drop across a 72-GPU rack behave differently from those across a single compute tray. Also, a single partially liquid-cooled server behaves very differently from a fully air-cooled or fully liquid-cooled server.\n\nThese days, AI infrastructure demands system-level validation rather than component-level checks, so the old chambers that would test up to a single 4U server are just not big enough. When we visited in June 2026 this was completed as you can see above, but in April 2026 we got to see it just before it was finished.\n\n## ASUS Extreme Environmental Chamber\n\nThe extreme environmental chamber handles conditions that exceed normal data center operation. This facility accepts entire racks and simulates temperatures from minus 40 degrees Celsius to 85 degrees Celsius. Humidity testing ranges from 10 percent (like some days here in Scottsdale, Arizona) to 98 percent (Taipei can often get to 98% relative humidity), covering edge deployments from cold outdoor installations to tropical environments with minimal climate control. Or, basically, where I started and ended my trip and where I ended up 14 hours later.\n\nOperating at these extremes serves two purposes. First, it validates that rack components survive shipping and storage in unconditioned spaces. That may seem small, but when the industry started shipping full NVIDIA GB200 NVL72 racks via air/ sea freight, temperature started to matter in the assembled systems. Many NVIDIA GB200 NVL72 and newer racks are shipped in climate-controlled trucks. Second, it identifies failure modes that emerge only under sustained thermal stress. Component aging accelerates at high temperatures, and running servers at 85 degrees Celsius for extended periods reveals weaknesses that would take years to surface in normal operation.\n\nASUS was operating the rack liquid-cooling system at 20 degrees Celsius during our visit, but was also testing higher coolant temperatures. Running warmer coolant reduces chiller energy consumption, though it can come at the cost of component reliability and performance headroom. Finding the right balance requires measuring actual failure rates across thousands of hours of accelerated aging tests.\n\nNoise levels inside this chamber are high when the environmental systems are running at full capacity. During our April 2026 visit, we walked around the chamber. It is huge, as you can see in the interior here, but the back houses all the power and fluid feeds, along with the environmental systems.\n\nRemote monitoring becomes essential during long test runs because spending hours in the chamber environment is impractical for technicians. Said another way, if you do not want to hear how loud the racks are and you want to keep the chamber operating its intended duty cycle, you need to automate the testing so the door stays shut.\n\n## ASUS AI System R&D Lab\n\nThe R&D lab focuses on rack-scale liquid-cooling integration rather than individual-server validation.\n\nRows of racks occupy this space, with power distribution running overhead and liquid-cooling infrastructure below the raised floor.\n\nThis arrangement mirrors what customers will deploy in production data centers, allowing ASUS to validate installation procedures and maintenance workflows before systems ship.\n\nLooking beneath the raised floor reveals the piping network that distributes coolant throughout the lab. Each rack connects to this manifold system, and the layout demonstrates how liquid cooling scales from single racks to multi-rack pods. Pipe diameter, valve placement, and leak detection all require coordination across the entire installation. ASUS also needs to verify that sensors work and appear reliably in its management software.\n\nThese days, the industry is tackling challenges beyond just detecting a leak, including ensuring a leak has minimal impact on the rack and the entire cluster’s operation.\n\nThis facility handles up to 1.1 megawatts of power across its rack positions. Liquid-cooled AI racks concentrate far more power density than traditional enterprise servers, and validating thermal performance at this scale requires infrastructure that matches production deployments.\n\nAir-cooled testing facilities cannot replicate the coolant distribution challenges that emerge at these power levels.\n\nASUS’s NVIDIA GB300 NVL72 rack integration differs fundamentally from testing individual NVIDIA GB300 nodes, shifting focus from thermal management of a single compute tray to coordinating cooling across 72 GPUs, managing coolant flow through multiple cold plates, and ensuring that leak detection and monitoring systems function correctly at rack scale. This lab exists to validate those system-level behaviors before customers deploy production hardware.\n\nA fun part of this lab is that the company is doing a lot of the work that many do not think about when they see a finished rack. We often see complete NVIDIA GB300 NVL72 racks, but we do not see validation happening with a few nodes hooked up to a Tektronix scope. That work happens here.\n\nOne neat aspect of the lab is that it uses a raised floor, but modern server gear weighs quite a bit. So there is a ramp built alongside it to offer another way to get gear up to the raised floor from the staging area.\n\nSince the lab has to handle both liquid-cooled and air-cooled components, it has air handlers like a traditional data center that cool the warm air that is contained in the hot aisle and brought down from the cieling, chilled, and then the cooler air is released back to the cold aisle.\n\nEar protection is important in the lab because, like a data center, it can get noisy.\n\nNext to the R&D lab is the power and chiller room for the setup.\n\n## ASUS R&D Lab Power and Chiller Room\n\nNext, we got to see the chiller room, which had various liquid cooling loops, the chiller, and the main power panels for the lab.\n\nGiant electrical panels line one wall of the chiller plant room, supporting the 1.1 megawatt power capacity of the testing facility. These panels distribute power to individual rack positions with monitoring at each branch circuit.\n\nA massive chiller occupies the other side of this room, handling heat removal from the liquid-cooling loops. A large chiller unit, visible during our visit, provides up to 1.3 megawatts of cooling capacity, exceeding the lab’s power draw, ensuring adequate headroom during peak test conditions. Liquid-cooled AI racks generate far more concentrated heat than air-cooled systems, requiring chillers sized for sustained high thermal loads since you cannot just buffer a short-duration burst with more air capacity like you can in an air-cooled system.\n\nSuch chiller design reduces power consumption by approximately 40 percent compared to conventional data center cooling approaches. Efficiency gains come from matching coolant temperature to actual server requirements rather than over-chilling for worst-case scenarios. When we say “chilled” these days, most data centers are using fluid closer to room temperature, and there is a trend towards using warm water. Really, what these liquid-cooled systems are after is a temperature differential and flow rate combination to cool racks. The warmer the water you can run while still achieving the necessary temperature differential, the less energy you spend cooling the fluid. Variable-speed compressors and intelligent load management adjust cooling output in response to real-time rack demand.\n\nEnergy efficiency at this scale translates directly into lower operational costs for customers running similar infrastructure in production environments.\n\n## ASUS Control Center Management\n\nRack-scale AI systems require monitoring tools that operate at multiple levels of granularity. ASUS Control Center provides visibility from individual nodes up through complete data center pods. During our visit, we saw how operators can click into a specific pod, drill down to a rack, and view telemetry for each node, including fan status, temperatures, and power draw.\n\nThis R&D lab runs its own telemetry system independent of the production management tools. Test cycles often continue for multiple days or weeks without interruption, requiring 24/7 monitoring to catch failures as they occur. Automated alerts notify technicians when parameters drift outside acceptable ranges, allowing rapid response before damage propagates through expensive AI hardware.\n\nDeployed clusters use the full ASUS Control Center, including Data Center Edition features, for operations management. Its interface aggregates data from multiple racks and presents it in a format that operations teams can act on quickly. Fan errors, coolant leaks, or thermal anomalies all trigger alerts with enough context to allow remote diagnosis.\n\nAI server facilities generate substantial acoustic output even with liquid cooling handling the majority of heat removal. Remaining air-cooled components, including power supplies and network switches, still require airflow, and large data centers accumulate noise from hundreds of fans across many racks. Detailed remote monitoring reduces the time technicians spend on the floor identifying issues, allowing more work to be done from control rooms or off-site locations.\n\nThe goal is to have reliable servers that you can verify without going into the data hall. Also, ensuring you can quickly see a failure about to happen or that has happened helps keep the entire cluster reliable. ASUS also takes its management platforms and has the AIDC or ASUS Infrastructure Deployment Center software that can monitor and manage systems, but also deploy applications across the cluster.\n\n## Final Words\n\nASUS’s thermal testing lab demonstrates how rack-scale validation differs from traditional server qualification. Environmental chambers that accept full racks, liquid-cooling infrastructure sized for multi-megawatt loads, and monitoring systems that operate at the node and rack levels all reflect the reality that AI servers deploy as integrated systems rather than standalone units.\n\nNVIDIA GB200 NVL72, NVIDIA GB300 NVL72, and NVIDIA Vera Rubin NVL72 integration introduces complexity that extends beyond individual compute trays. Coolant distribution across 72 GPUs, CPUs, networking, storage, power management at rack scale, and leak detection in production environments all require validation before customers commit to deployments costing millions of dollars. Facilities like this one allow vendors to identify and resolve issues during development rather than after systems reach customer data centers.\n\nThis rack-scale approach reflects how the industry is adapting to liquid-cooled AI infrastructure, and it is becoming the price of admission for delivering these high-end systems. Rack-scale integration remains central to successful deployments as vendors continue refining their testing methodologies. We always love getting to see these labs since they are usually a vital part of the server systems we test, but we oftentimes do not get to tour them and show you.", "url": "https://wpnews.pro/news/asus-thermal-lab-tour-2026-testing-ai-servers", "canonical_source": "https://www.servethehome.com/asus-thermal-lab-tour-2026-testing-ai-servers/", "published_at": "2026-07-10 17:00:46+00:00", "updated_at": "2026-07-10 17:16:49.249701+00:00", "lang": "en", "topics": ["ai-infrastructure", "ai-chips", "ai-products"], "entities": ["ASUS", "NVIDIA", "HGX B200", "GB200 NVL72", "Taiwan"], "alternates": {"html": "https://wpnews.pro/news/asus-thermal-lab-tour-2026-testing-ai-servers", "markdown": "https://wpnews.pro/news/asus-thermal-lab-tour-2026-testing-ai-servers.md", "text": "https://wpnews.pro/news/asus-thermal-lab-tour-2026-testing-ai-servers.txt", "jsonld": "https://wpnews.pro/news/asus-thermal-lab-tour-2026-testing-ai-servers.jsonld"}}