The data center industry is facing a convergence of multiple pressures. Global demand for compute power, combined with heavy AI workloads, cloud expansion, and edge deployments, is straining the limits of conventional land-based infrastructure. 

Data center cooling accounts for roughly 40% of a typical energy consumption. Water usage at large hyperscale campuses runs into hundreds of thousands of gallons per day. Power grid constraints are delaying new builds by years in some markets. And communities near data center clusters are pushing back harder than ever.

Amidst this swarm of pressures, the far-reaching idea of putting data centers in the ocean comes as a possible solution.

Underwater data centers, also referred to as submerged data centers or subsea data centers, are sealed, pressure-tolerant computing modules deployed on the ocean floor or other bodies of water. The concept uses the natural thermal mass of the sea as a free, continuous cooling medium, dramatically reducing or eliminating the need for conventional air conditioning and freshwater cooling systems.

A new frontier that was once just an idea but is no longer a theoretical exercise. Microsoft ran a proven subsea system off the coast of Scotland for two years. China has launched what is widely regarded as the world's first commercial underwater data center. 

Startups like Subsea Cloud are commercializing the technology at scale. And the intersection of surging AI compute demand with tightening environmental regulation is making the economics increasingly compelling.

How Underwater Data Centers Work

An underwater data center is a sealed computing facility placed beneath the surface of a body of water, such as an ocean or river. It hosts servers and related equipment and performs the same tasks as a land-based data center, while using the underwater environment to help manage cooling, power use, and space limitations.

The principle is straightforward but powerful. Deep ocean temperatures are stable, typically ranging between 2°C and 15°C depending on depth and geography. Rather than burning electricity to drive compressors and chillers, an underwater data center passively exchanges heat with the surrounding seawater.

This alone can dramatically reduce a facility's Power Usage Effectiveness (PUE), the ratio of total energy consumed to the energy used by computing equipment, moving it well below the industry average of around 1.5 toward near-ideal values of 1.0 to 1.1.

Nitrogen Environments and Sealed Operations

Microsoft's Project Natick data center featured 864 servers and 27.6 petabytes of storage in a cylinder filled with unreactive nitrogen gas, tethered to land by a cable that provided both a fiber connection and power. The nitrogen atmosphere is a key engineering choice: it eliminates oxygen-based corrosion, one of the primary failure modes for electronic components in conventional facilities. 

Without human traffic, vibration from people walking the floors, and the repeated thermal cycling caused by HVAC systems powering up and down, hardware components experience fundamentally different, and in many respects more benign, operating conditions.

Connectivity, Power Delivery, and Modularity

Power and data connectivity are delivered via umbilical cables from shore or from offshore renewable installations. The modularity of sealed pod or cylinder designs means capacity can theoretically be scaled by deploying additional units, much like stacking shipping containers. 

Projects like Subsea Cloud's pressure-equalized pods offer simpler deployment, easier maintenance, and strong security, pointing toward a future where standard containerized modules can be deployed, serviced, and recovered with relatively low operational friction.

Proximity to Coastal Populations

One underappreciated advantage of the undersea approach is geographic positioning. More than half the world's population lives within 120 miles of the coast, which means underwater data centers deployed on the continental shelf can serve as ultra-low-latency edge nodes for a massive addressable user base without consuming any land in already-crowded coastal urban areas.

From Experiment to Commercial Reality

Microsoft Project Natick

No discussion of underwater data centers is complete without Microsoft's Project Natick, the most rigorously documented subsea data center experiment to date.

Source: Microsoft News

Mini Case Study: Microsoft Project Natick

Microsoft's first subsea deployment occurred in 2016 off the California coast for a three-month trial run. The second and far more significant phase launched in 2018, when a sealed container with 12 racks rested on the ocean floor approximately 117 feet deep off Scotland's Orkney Islands. The facility operated for two full years before being retrieved in 2020.

The headline finding was remarkable. Microsoft plans to study why the Natick servers proved eight times more reliable than those in a replica set up on land with stable temperatures, no oxygen corrosion, and no people to bump and jostle components, which is thought to be the reason.

Eight times more reliable. For data center owners and operators, this single data point deserves serious attention. If hardware failure rates can be reduced by an order of magnitude, the implications ripple across capital expenditure planning, sparing strategy, service level agreements, and total cost of ownership models.

The project also surfaced important findings around energy resilience. Natick relied on wind, solar, and experimental tidal technologies, and Spencer Fowers of Microsoft's Special Projects said, "We have been able to run really well on what most land-based data centers consider an unreliable grid. We are hopeful that we can look at our findings and say maybe we don't need to have quite as much infrastructure focused on power and reliability."

Looking ahead, Microsoft noted that if it were to create a data center with the same capabilities as a standard Microsoft Azure deployment, it would require dozens of the vessels, and it could be used for edge deployments or for data that needs to be highly secured. William Chappell, vice president of mission systems for Azure, specifically highlighted tests of post-quantum encryption technology as a particularly compelling use case for the isolated, physically secure underwater environment.

China Moves from Experiment to Commercial Operation

While Microsoft's Natick remained a research program, China has moved aggressively toward commercialization. China is pulling ahead of the rest of the world in sinking data centers that power AI into the ocean as an alternate way to keep them cool, as it bets big on artificial intelligence, cloud computing, and other digital technology.

China has officially launched the world's first commercial underwater data center, in an effort to lower the amount of energy needed to keep servers cool. The facility, developed by Hainan-based Shanghai Hailanyun Technology, was initially piloted off the Hainan coast. 

China's Hainan project provides large-scale underwater AI computing with minimal environmental impact, and a more advanced, larger-scale installation is now being developed off Shanghai.

The strategic logic for China is clear: these massive collections of servers gobble up growing amounts of energy, and each one cycles through hundreds of thousands of gallons of water a day to carry away the heat they generate. 

Moving infrastructure underwater addresses both the energy and freshwater constraints simultaneously, a dual benefit that is particularly valuable in a country where many inland regions face water scarcity.

The Broader Competitive Landscape

Beyond Microsoft and China, the field now includes several notable players:

Nautilus Data Technologies operates barge-based data centers with efficient closed-loop cooling systems. Subsea Cloud has developed pressure-equalized pods offering simpler deployment and maintenance. 

Even Google explored floating data center concepts, though that program has since been retired. These projects differ in design and scale, but all share the goal of rethinking where and how data infrastructure can be deployed.

The convergence of proven engineering results, a first commercial deployment, and multiple well-funded entrants signals that submerged data centers are transitioning from curiosity to a recognized segment of the infrastructure market.

The Environmental Debate, Regulatory Challenges, and Pushback

The Environmental Promise

The headline environmental claim for underwater data centers rests on three pillars: reduced energy consumption through passive seawater cooling, elimination of freshwater cooling withdrawal, and the potential to co-locate with offshore renewable energy sources such as wind and tidal generation. 

If these benefits are realized at scale, submerged data centers could represent a genuinely lower-impact model of digital infrastructure relative to conventional hyperscale campuses.

Data centers currently use up around 480 terrawatts of electricity as of 2025, and cooling is the dominant variable cost. Any technology that substantially reduces that figure has significant implications for carbon accounting, grid planning, and corporate sustainability commitments.

Environmental Concerns and Ecological Risks

Despite the optimistic framing, underwater data centers face serious and legitimate environmental scrutiny.

The primary concern is thermal pollution. Large-scale heat rejection into a contained marine environment, a bay, estuary, or shallow coastal zone, could elevate local water temperatures, stressing coral reefs, altering fish spawning behavior, and shifting species distributions. 

Unlike air-cooled facilities that discharge heat into the atmosphere, ocean-cooled facilities transfer thermal load directly into an ecosystem with finite capacity for absorption.

Seawater cooling cuts energy use but raises ecological and maintenance concerns. Biofouling, the accumulation of marine organisms on exterior surfaces, is a recognized challenge for any submerged infrastructure. Chemical antifouling treatments carry their own toxicity risks. The physical disturbance of seabed habitats during installation and retrieval is another concern that requires proper environmental impact assessment.

There are also concerns about chemical contamination from coolant leaks, lubricants, and other industrial fluids within sealed modules. While the sealed nitrogen-filled designs minimize many of these risks under normal operating conditions, catastrophic failure scenarios and end-of-life decommissioning require careful planning.

Community and Legislative Pushback on Data Center Expansion

The broader backlash against data center development provides essential context for understanding why unconventional locations, including the ocean floor, are gaining traction. Energy, water, land use, noise, and e-waste are all sleeping giants in terms of community concern, but communities near data centers are increasingly waking up.

In the United States, there has been notable pushback at the federal legislative level. Senator Bernie Sanders and colleagues have raised concerns about the resource implications of large-scale AI infrastructure build-out, including data centers' growing share of national electricity demand and water consumption. This legislative pressure reinforces the incentive for the industry to explore lower-impact models.

The question of whether underwater data centers are truly practical remains open, and that question is being answered not just by engineering benchmarks but by the regulatory and community reception that conventional land-based facilities are increasingly facing.

Maintenance, Access, and Operational Realities

A frequently raised practical objection to submerged data centers is the challenge of maintenance. When hardware fails in a conventional facility, a technician walks to the rack. When hardware fails in a sealed module on the ocean floor, the options are limited to remote management, waiting for a scheduled retrieval, or mounting a complex and expensive recovery operation.

Microsoft's Project Natick was designed with this constraint in mind; the premise was that hardware reliability would be high enough that human intervention would be rarely necessary. The eight-times improvement in server reliability documented at Natick suggests this bet may be well-placed. 

However, it remains untested at the scale and density of a full production hyperscale deployment. The economics of maintenance access need to be carefully modeled before large-scale capital commitments.

Conclusion: What Data Center Professionals and Investors Need to Do Next

Underwater data centers sit at an inflection point. The technology has moved beyond proof-of-concept. Commercial operations are live. The macroeconomic drivers, AI-driven compute demand, grid constraints, freshwater scarcity, rising land costs, and tightening environmental regulation, are all accelerating, not abating.

For data center developers and owners, the immediate priority is monitoring the Chinese commercial deployment closely and engaging with vendors like Subsea Cloud and Nautilus to understand where modular underwater or near-water infrastructure fits within your capacity expansion roadmap. 

The use cases most immediately compelling are edge deployments in coastal markets, high-security applications, and facilities co-located with offshore renewable energy installations.

For working professionals in the data center industry, engineers, facilities managers, and operations teams, the Project Natick reliability data is a genuine signal worth taking seriously. Understanding nitrogen-atmosphere environments, seawater heat exchange systems, and remote monitoring architectures will become increasingly relevant skills as this segment matures.

For businesses and investors evaluating where to allocate capital or strategic attention within digital infrastructure, the underwater and near-water data center segment represents a high-conviction emerging market. The convergence of proven technical performance, first commercial operations, clear demand drivers, and favorable differentiation from the regulatory and community headwinds facing conventional campuses creates a compelling investment thesis.

The deeper question is not whether underwater data centers will become part of the infrastructure landscape; they already are. The question is how fast the segment scales, which deployment models prove most economically durable, and how the industry navigates the genuine environmental responsibilities that come with putting critical infrastructure into the ocean.