Nautilus Data Technologies has announced that its EcoCore® Facility Cooling Distribution Unit (FCD) is now recognized as NVIDIA-Validated AI Factory Infrastructure. Listed on NVIDIA's DSX Infrastructure Marketplace, the unit meets rigorous functional requirements for AI factory and data center deployments. This validation positions Nautilus as a key infrastructure partner for operators and hyperscalers managing next-generation AI GPU deployments requiring high-performance, liquid-to-liquid cooling solutions.
NVIDIA DSX Marketplace Infrastructure Recognition
The EcoCore® FCD has achieved a significant milestone by becoming the first liquid-to-liquid cooling distribution unit to receive Large-Scale CDU validation. It currently holds the highest-rated capacity among all liquid-to-liquid systems listed in the NVIDIA DSX Infrastructure directory. This recognition follows independent performance testing completed in May 2026, where the unit successfully met over one hundred specific NVIDIA technical requirements. By appearing in the NVIDIA-Validated AI Factory Infrastructure marketplace, the EcoCore® FCD provides a verified option for decision-makers seeking to align their cooling, power, and infrastructure with NVIDIA's specific functional standards for intensive AI workloads and large-scale GPU environments.
EcoCore FCD Technical Performance Metrics
Technical validation highlights the EcoCore® FCD's superior thermal management capabilities compared to industry alternatives. In N+1 mode, the unit provides 3.2 MW of cooling capacity, while N-mode dual-pump operation demonstrated capacity exceeding 4.1 MW. A critical performance differentiator is the unit's 2°C approach temperature difference (ATD), which is less than half the 4.8°C ATD reported by competing systems. The architecture is water-source agnostic, supporting process water, chilled water, river water, seawater, or wastewater. Additionally, the design integrates with Nautilus's Air Detection Removal System degassing technology and allows for in-place maintenance and upgrades, supporting long-term resilience for evolving AI computing requirements.
Key Takeaways
- The EcoCore® FCD is the first liquid-to-liquid CDU to achieve Large-Scale CDU validation from NVIDIA.
- The unit demonstrated a 2°C approach temperature difference (ATD), significantly lower than the 4.8°C ATD of competitors.
- EcoCore® FCD offers a cooling capacity exceeding 4.1 MW in N-mode dual-pump operation.
TechInsyte's Take
In our view, this validation signals a shift toward specialized, high-density cooling requirements that generic data center infrastructure can no longer meet. By securing NVIDIA's validation, Nautilus is moving beyond mere hardware supply to become a critical component of the AI compute stack. The ability to maintain a 2°C ATD while remaining water-source agnostic suggests that Nautilus is targeting the specific thermal bottlenecks that currently limit hyperscale AI expansion. This positioning is strategic for operators facing extreme power densities.
Questions & Answers
How does the EcoCore FCD compare to competing liquid-to-liquid cooling systems in terms of thermal efficiency?
The EcoCore FCD demonstrates a 2°C approach temperature difference (ATD), which is notably less than half of the 4.8°C ATD reported for competing systems, indicating superior thermal management.
What specific capacity metrics were validated for the EcoCore FCD during NVIDIA testing?
The unit demonstrated a cooling capacity exceeding 4.1 MW in N-mode dual-pump operation and 3.2 MW in N+1 mode.
What operational flexibility does the EcoCore FCD offer regarding water sources?
The unit is water-source agnostic, meaning it can operate using process water, chilled water, river water, seawater, or wastewater.
How does the NVIDIA validation impact the procurement process for data center operators?
The inclusion in NVIDIA's DSX Infrastructure Marketplace allows operators and hyperscalers to identify validated partners that meet specific functional requirements for next-generation AI GPU deployments.
Source: PRNEWSWIRE
