Nautilus Data Technologies has been recognized as NVIDIA-Validated AI Factory Infrastructure after its EcoCore Facility Cooling Distribution Unit was added to NVIDIA’s DSX Infrastructure Marketplace.
The recognition makes the EcoCore FCD available to data center operators and hyperscalers looking for cooling equipment validated for AI factories and high-density GPU deployments.
According to Nautilus, the system has the highest-rated capacity among the liquid-to-liquid cooling distribution units listed in the NVIDIA directory. It is also the first product in the category to receive Large-Scale CDU validation.
The development matters because the growth of AI computing is creating much greater cooling requirements than conventional data center workloads.
As companies deploy larger GPU clusters, cooling infrastructure must handle more heat while maintaining the temperatures and flow rates required by increasingly powerful hardware.
EcoCore FCD Delivers More Than 4.1 MW Of Cooling
The EcoCore FCD completed testing against more than 100 NVIDIA technical requirements.
During testing, the unit delivered more than 4.1 megawatts of cooling capacity while operating with both pumps in N-mode. It also reached 3.2 megawatts in N+1 mode, where one pump remains available as a backup if another fails.
That redundancy is important for AI data centers because a cooling system failure can force operators to reduce computing performance or shut down equipment to prevent overheating.
Nautilus also reported an approach temperature difference of 2°C. The company said that was less than half the 4.8°C reported for competing systems.
Approach temperature difference measures the gap between the temperature of the facility water entering the cooling unit and the temperature of the water supplied to the computing equipment.
A smaller difference generally means the system can transfer heat more efficiently, although real-world performance will still depend on facility design, operating conditions and the equipment connected to it.
NVIDIA Marketplace Highlights Validated Infrastructure
NVIDIA’s AI factory marketplace is designed to help organizations find infrastructure products that meet the company’s requirements for AI deployments.
Those products include the cooling, power, networking and computing systems needed to build facilities around high-density GPU clusters.
NVIDIA says validated designs can reduce deployment complexity, improve compatibility between components and help operators achieve more predictable results.
For Nautilus, appearing in the marketplace gives potential buyers another way to compare the EcoCore FCD with other cooling distribution units intended for large AI installations.
The validation follows independent performance results Nautilus announced in May 2026, which examined the unit’s ability to meet pressure, temperature and flow requirements associated with current and upcoming GPU systems.
Liquid Cooling Becomes More Important For AI Factories
Traditional data centers have relied heavily on air cooling, where fans and air-conditioning systems remove heat from server rooms.
That becomes more difficult as racks contain more GPUs and consume more power.
Liquid cooling moves heat away from computing equipment using water or another fluid. Because liquids can carry more heat than air, these systems can support denser hardware configurations without requiring the same amount of airflow.
The EcoCore FCD sits between a facility’s water system and the cooling loop connected to the computing equipment. Its job is to regulate water flow, temperature and pressure while keeping the two sides of the system separated.
Nautilus says the unit can operate with process water, chilled water, river water, seawater or wastewater, giving operators more flexibility when designing facilities in different locations.
The company also says the system can be maintained and upgraded in place. It can connect with Nautilus’ Air Detection Removal System, which removes trapped gases that could reduce cooling performance or damage equipment.
Nautilus Targets Faster AI Data Center Deployments
Nautilus says the EcoCore FCD is available globally with delivery times of around 15 weeks from the date of an order.
The company also reports that its cooling systems have accumulated more than 650,000 unit-hours of production operation across three generations of products.
Nautilus CEO Rob Pfleging said the NVIDIA recognition supports the company’s goal of creating cooling infrastructure for both current and future AI workloads.
The company is targeting AI developers, hyperscale operators, enterprises and colocation providers building facilities for high-density computing.
The NVIDIA listing does not guarantee how the EcoCore FCD will perform in every data center. Actual results will depend on workload density, facility water conditions, redundancy requirements and the design of the broader cooling system.
However, completing NVIDIA’s validation process gives operators another technical reference when choosing equipment for large GPU deployments.
As future AI platforms consume more power and generate more heat, cooling systems will increasingly influence how quickly new facilities can be built, how densely hardware can be installed and how reliably those systems can operate.
For Nautilus, the EcoCore FCD’s addition to NVIDIA’s marketplace places the company closer to that growing part of the AI infrastructure market.
Via: JSA | Nautilus Data Technologies | PR Newswire | NVIDIA
