Underwater AI Data Centers: Benefits, Risks & Future

Could the Ocean Be the Next Home for AI Data Centers?

As artificial intelligence is evolving swiftly, not only are the tech nerds hungry for its breakthroughs and evolution. AI is also hungry. And not just for data. This time it is hungry for compute, power, cooling, and space.

The artificial intelligence boom is driving unprecedented demand for data centers, raising concerns about rising energy consumption, water use, and carbon footprints. These challenges are making firms look beyond outdated land-based data centers.

Some developers are now exploring the ocean as a new location for AI infrastructure in the hopes that underwater data centers could improve energy use and cooling efficiency while using less fresh water and land area than onshore buildings.

But shifting servers into the ocean does not make other underlying environmental hurdles such as energy consumption and carbon emissions disappear. And it creates new concerns about harm to the marine environment, as well as questions about how these data centers can be regulated.

Key Takeaways: Underwater AI Data Centers

Growing AI Infrastructure Demand: Artificial intelligence requires increasing amounts of computing power, electricity, cooling capacity, and physical infrastructure.
Underwater Data Centers: These facilities place sealed computing equipment beneath the ocean's surface to explore alternative cooling and deployment methods.
Microsoft Project Natick: Microsoft's underwater data center experiment near Scotland contained 864 servers and demonstrated promising hardware reliability results.
Potential Cooling Benefits: Seawater can absorb heat from computing equipment, potentially reducing dependence on some conventional cooling systems.
Reduced Freshwater Requirements: Certain underwater cooling designs may reduce the need for freshwater-based cooling.
Maintenance Challenges: Repairing or replacing failed underwater equipment can require retrieving an entire sealed module.
Environmental Risks: Waste heat, installation activities, and marine infrastructure could affect surrounding ecosystems.
Renewable Energy Opportunities: Offshore wind and other renewable energy sources could potentially support ocean-based computing facilities.
Commercial Limitations: Underwater infrastructure faces challenges involving corrosion, pressure, networking, scalability, maintenance, and environmental permits.
Future Outlook: Ocean-based data centers may become a specialized infrastructure option, but they are unlikely to replace conventional data centers in the near future.

Quick Answer: Can AI Data Centers Be Built Underwater?

Yes, underwater data centers are technically possible. Microsoft's Project Natick successfully demonstrated the operation of a sealed data center module on the seabed near Scotland. The experiment reported substantially lower hardware failure rates than a comparable group of land-based servers.

However, underwater data centers are not yet a proven mainstream solution for large-scale AI computing. Challenges include equipment maintenance, environmental impacts, energy supply, network connectivity, regulatory approval, and the difficulty of upgrading hardware.

The ocean could support some future computing infrastructure, but commercial adoption will depend on whether underwater facilities can operate reliably, economically, and sustainably.

The Rise of Ocean-based AI Infrastructure

Almost a decade ago, Microsoft launched a research project to explore the feasibility, benefits, and challenges of underwater data centers. Part of that effort included setting up a waterproof data center on the seafloor near the Scottish islands. It contained 864 servers and was connected to shore by an underwater cable.

After two years, Microsoft reported that the servers in the underwater data center failed at about 1/8th the rate of servers in comparable land-based data centers. The company is still studying the possible reasons but hypothesizes that in a sealed underwater environment the equipment is less exposed to oxygen, humidity, and temperature fluctuations.

However, Microsoft later ended the project in 2024 and chose not to build more underwater data centers. The company didn’t say why, but other analyses suggest the reasons could include regulatory issues, including the need for environmental permits, as well as a wish for quick upgrades and replacements for the computer equipment inside.

Instead, the company has concentrated on land-based data centers, which can be larger and easier to expand, and also easier to access to repair or replace equipment.

The Promise of Ocean-based Data Centers

These multiple approaches deliver ways to lessen demand for grid-supplied electricity for powering data center computers and cooling equipment, as well as using less fresh water.

The distance from people’s homes could also be a benefit for data centers in or on the ocean. A Gallup poll in March 2026 found that 60% of Americans were not in favor of building AI data centers in their communities. However, more than half of the world’s population lives within 120 miles of a coast. Underwater could be another way to keep data centers physically close to users for speedy service.

Maintenance, though, is a big challenge. If a computer fails underwater, it cannot be repaired or replaced on site. The whole sealed data center module may need to be brought to the surface, even if just one computer requires work.

1. Natural Cooling Using Seawater

Cooling is an essential part of data center operations. Servers generate heat while processing information, and excessive temperatures can reduce performance or damage equipment.

Conventional data centers use combinations of air cooling, liquid cooling, chillers, heat exchangers, and other thermal management technologies.

Underwater facilities can potentially transfer heat to surrounding seawater through specialized heat-exchange systems.

This approach may reduce the need for certain mechanical cooling components, depending on the design and environmental conditions.

However, seawater cooling is not energy-free. Pumps, heat exchangers, monitoring systems, and supporting infrastructure may still consume electricity.

2. Reduced Freshwater Consumption

Some conventional data centers use evaporative cooling systems that consume freshwater.

In regions experiencing water scarcity, this can create additional pressure on local resources.

Underwater data centers may reduce freshwater requirements by using seawater as a heat-transfer medium.

The actual water savings depend on the facility's cooling design and the land-based system used for comparison.

3. Reduced Demand for Land

Large data centers require space for server buildings, electrical infrastructure, cooling systems, and supporting facilities.

In densely populated regions, land availability and construction costs can become significant challenges.

Underwater modules could reduce the amount of land needed for some computing infrastructure.

However, these projects would still require onshore power connections, networking facilities, monitoring equipment, and maintenance infrastructure.

4. Potential Proximity to Coastal Users

A significant portion of the world's population lives in coastal regions.

Deploying computing infrastructure near these populations could potentially reduce the distance data travels between users and servers.

This may be relevant for applications that require low latency, such as real-time analytics, industrial monitoring, and some AI inference services.

Nevertheless, network routing and connectivity quality are often more important than physical distance alone.

5. Opportunities for Renewable Energy Integration

Ocean-based data centers could potentially be connected to offshore wind farms or other renewable electricity sources.

Locating computing infrastructure near renewable generation facilities may create opportunities for alternative power-delivery designs.

However, renewable electricity production varies over time. Data centers generally require continuous power, so reliable grid connections, storage, or other backup arrangements would still be necessary.

Underwater vs Traditional Data Centers: Key Differences

Factor	Underwater Data Centers	Traditional Land-Based Data Centers
Location	Seabed or marine environment	Land-based buildings
Cooling	Potential use of surrounding seawater	Air, liquid, evaporative, or hybrid cooling
Freshwater Use	Potentially lower for certain designs	Varies significantly by cooling technology
Maintenance	Physical access is difficult	Equipment is generally accessible
Hardware Upgrades	May require module retrieval	Usually easier to perform
Land Requirements	Reduced facility footprint on land	Requires dedicated physical space
Environmental Risks	Marine heat discharge and installation impacts	Land use, water consumption, emissions, and heat
Commercial Maturity	Experimental and specialized deployments	Established global industry
AI Hardware Replacement	Potentially complicated	Generally more practical

Can the Ocean Sustain AI?

One of the major environmental concerns about ocean-based data centers consists of seawater used for cooling. Discharging warm or hot water can potentially affect oxygen levels, pH, and marine life in the water.

The heat is already apparent at the few seaborne data centers now operating. Several marine species rely on stable water temperature for breeding, feeding, and migration, highlighting concerns that heat released by various underwater data centers could create localized thermal pollution and change marine ecosystems. And the ocean is already under pressure. UNESCO, including in conservation, estimates that about 50% of marine ecosystems are already degraded or used unsustainably.

The ocean is already warming along with the atmosphere, without extra waste heat from data centers. The additional heat is threatening coral reef and mangrove ecosystems, seagrasses, and other aspects of the marine food web. As that warming continues, ocean water will no longer be useful for cooling electronic equipment in some regions.

Thus, underwater data centers could assist AI growth while easing some of the pressure on land, energy, and water. But the real test is whether the ocean can become AI’s next computing frontier without becoming its next environmental issue.

Understanding the Future of Ocean-Based AI Data Centers

The idea of setting up data centers underwater sounds promising, but turning it into a practical solution for the growing demand for AI computing will take time. Firms must find ways to manage computing performance, energy efficiency, and environmental sources, like offshore wind and tidal power, to support their operations.

With the right infrastructure, these facilities could assist in decreasing dependence on conventional electricity sources while taking advantage of naturally cool seawater for heat management. However, renewable energy alone will not solve each challenge. Underwater data centers still need reliable power connections, specialized equipment, secure communication networks, and systems that can withstand corrosion, water pressure, and harsh marine conditions.

Environmental monitoring will be equally crucial. Before increasing these projects, developers need to understand how heat discharge, underwater structures, and installation activities could affect nearby marine ecosystems. Careful site selection, strict environmental assessments, and regular monitoring could help overcome these risks.

Conclusion

The increasing demand for AI is pushing the technology industry to ruminate on where and how computing infrastructure should be built. From increasing electricity consumption to cooling needs and limited land availability, traditional data centers are facing challenges that need new approaches.

Ocean-based data centers deliver an interesting possibility. They could take advantage of natural cooling, decrease the need for certain land-based resources, and bring computing infrastructure closer to coastal populations. However, these advantages are not guaranteed, and underwater operations introduce their own challenges, including costlier maintenance, environmental risks, and regulatory complications.

The actual question is not whether we can place AI data centers underwater, but whether we can operate them effectively and responsibly over the long term. That will need more research, practical testing, reliable renewable energy, and a transparent understanding of their impact on marine ecosystems.

As AI continues to rise, the ocean may become one of several locations considered for future computing infrastructure. But technological progress should not come at the cost of the environment. The next generation of data centers must do more than power smarter AI; it must also support a more sustainable digital future.