TL;DR: After running a fully liquid-cooled AI data center for years, Nautilus turned real-world experience into a product. EcoCore is an operator-built, fast-to-deploy liquid-cooling platform and CDU line designed to deliver reliable, AI-ready cooling to any data center.
For years, Nautilus operated one of the worldโs most advanced liquid-cooled data centers. That experience, along with 400,000+ unit hours of continuous operation (over 100K being direct-to-chip), taught us something that no amount of modeling or lab testing ever could: itโs one thing to design for performance, and another to live with your own technology every single day servicing customers. We saw what this technology could do and how important it was for the future of AI cooling.
That perspective changed everything.
Itโs the reason we shifted from being an operator to becoming a product company.
Why We Shifted to Liquid Cooling, And What It Means for Data Centers
When we built and ran our own facility, we proved what many thought was impossible: liquid cooling could support the most demanding AI and HPC workloads with unmatched efficiency, uptime, and sustainability. But we also saw a much bigger opportunity.
The industry didnโt just need another efficient data center, it needed scalable, ready-to-integrate infrastructure that any operator could deploy.
So instead of building more of our own facilities, we decided to make what weโd learned available to the market.
We turned operational insight into a product line, EcoCore, built to help others deploy AI-ready, liquid-cooled capacity in months instead of years.
From Operations to Engineering: The Proven Lessons Behind Our Liquid Cooling Platform
Operating one of the only 100% liquid-cooled AI facilities in the world gave us a front-row seat to every challenge real-world cooling faces: flow variability, pressure balance, filtration, redundancy, and the unexpected behavior of AI workloads under thermal stress.
Those lessons became our design principles.
The EcoCore CDUs reflect these principles:
- Leak-proof reliability through our variable pressure Leak Prevention System.
- Adaptive configurations that handle everything from direct-to-chip to rear-door cooling to immersion and to what comes next.
- Vendor-agnostic integration that fits into legacy or new-build environments.
- Rapid deployment, with units ready in ~12โ16 weeks
A Liquid Cooling System Purpose-Built by Operators for Operators
Our goal isnโt to sell cooling, itโs to solve the same problems we once had.
We know what downtime feels like. We know how hard it is to retrofit an AI-ready hall. And we know what it takes to deliver high-density performance safely, efficiently, and fast.
Thatโs why the shift from operator to liquid cooling provider wasnโt a departure from our roots.
It was a decision to scale them.
Because in an AI-driven world, experience isnโt just an advantage, itโs the foundation of every product we make.
Future-proof cooling, built by the people whoโve lived it. Discover EcoCore.
FAQs
Why did Nautilus move from operating a liquid-cooled data center to offering liquid cooling products?
Nautilus made the shift after running a fully liquid-cooled AI facility and learning firsthand what works and what fails. That real-world experience helped them design a product line to solve the problems they faced.
How did operating a liquid-cooled data center shape the design of the new cooling products?
Daily operational challenges around flow, pressure, reliability, and maintenance showed the team which engineering decisions actually matter. Those lessons were built directly into the product.
What makes these liquid cooling products useful for data center operators?
They are designed by people who have operated liquid-cooled systems themselves, so the products focus on reliability, easier deployment, cleaner integration, and practical day-to-day needs.
How quickly can these data center liquid-cooling systems be deployed?
EcoCore is designed for quick rollout, with CDUs typically ready in about 12-16 weeks, so operators can add liquid-cooled capacity much faster than traditional build cycles.
Why does real-world experience matter for liquid cooling in AI data centers?
Experience shows how systems behave under real load, not just in a lab. It helps prevent failures, improves uptime, and ensures the cooling system can handle the demands of high-density AI workloads.
