Cooling
Can Underwater Data Centers Reduce Cooling Costs?
Seawater can reduce mechanical cooling plant, but cooling savings are only one line in the operating budget. Materials, retrieval, power quality, and networking still decide the total.
· 9 min read
By Seabase Editorial
← Articles index · Home · Infrastructure · Nori · About
Topics: Cooling · Environmental
Yes, underwater data centers can reduce some cooling costs. No, immersion does not automatically make AI infrastructure cheap. The honest answer depends on thermal design, local seawater conditions, maintenance, and what you compare against.
Cooling can get cheaper underwater. The whole facility only gets cheaper if marine operations do not erase the savings.
How marine cooling works
Sealed modules reject heat through their surfaces or dedicated exchangers into surrounding water. Because seawater is a large moving thermal reservoir, designs can avoid some cooling towers, condenser plants, and land footprint associated with conventional campuses.
Internal IT cooling may still use air, cold plates, or direct liquid cooling with a CDU before heat reaches the seawater interface.
Where savings can appear
- Lower mechanical plant capital for large air-side systems
- Lower land area dedicated to cooling equipment
- Potentially lower water consumption versus evaporative towers
- Stable sink temperatures in some coastal environments
Where costs remain
- Pressure-tolerant enclosures and marine-grade materials
- Coatings, corrosion control, and biofouling management
- Installation and retrieval campaigns
- Monitoring for thermal and ecological thresholds
- Insurance, classification, and specialized crews
- Power conversion and cable losses to shore
A project can win on heat rejection and still lose on logistics.
Power usage effectiveness
PUE is a common efficiency ratio for data centers. It helps compare facility overhead, but it does not by itself describe total capital cost, land use, or cooling complexity. Glossary: PUE.
Seabase is designing toward approximately 1.05 PUE through direct liquid cooling and marine heat rejection, as a design target subject to engineering validation and operating conditions. See AI Infrastructure Statistics.
The environmental side of heat rejection
Moving heat into the ocean is not free of impact. Local temperature rise, mixing, and habitat interaction matter. Continuous monitoring is part of responsible design. See Environmental Accountability.
Bottom line
Underwater data centers can reduce cooling costs when seawater heat rejection displaces large mechanical plants and the site is otherwise well chosen. Evaluate the full stack: thermal design, power, fiber, maintenance, and environmental controls. Start from What Is a Subsea Data Center? and compare architectures in Underwater vs Land Data Centers.
Next step
Discuss thermal design and coastal deployment assumptions with Seabase.
Related research
- What Is a Subsea Data Center? (Complete Guide)
A complete guide to subsea and underwater data centers: history, benefits, challenges, power, networking, cooling, maintenance, environmental concerns, floating vs subsea, and AI inference.
- Environmental Accountability for Subsea AI Infrastructure
How Seabase measures, monitors, and reports on the environmental footprint of subsea AI compute infrastructure: thermal baselines, acoustic impact, material stewardship, seabed disturbance, and the limits of zero-impact claims.
- Underwater Data Centers vs Land Data Centers
Compare underwater and land data centers on cooling, land use, maintenance, power, networking, latency, cost, and environmental review. When each approach fits.
- Floating vs. Subsea Data Centers: Why Seabase Chose Subsea
Floating data centers can reduce land and cooling constraints, but remain exposed surface marine assets. Seabase explains why modular subsea infrastructure better fits its goals for metropolitan proximity, stable operation, depth-based cooling, serviceability, and global scale.
- Reducing the Community Footprint of AI Infrastructure
Seabase is developing modular coastal and subsea compute infrastructure intended to reduce the land, visual, acoustic, freshwater, and development footprint of high-density AI capacity.