en.Wedoany.com Reported - China has launched a commercial underwater data center project, a technological direction worth noting, but its industrial disruptiveness should not be overinterpreted.

Data center operations generate substantial heat, and the ocean, with its immense heat capacity, serves as a natural cooling medium. The concept of deploying sealed data center modules underwater, leveraging seawater for primary cooling while reducing land use and positioning computing facilities closer to coastal demand centers, is appealing.

This technology is not a fantasy. Microsoft's Project Natick deployed an underwater data center pod containing 12 racks and 864 servers off the Orkney Islands in Scotland in 2018, and retrieved it in 2020. Microsoft reported that the server failure rate in the underwater pod was one-eighth that of a land-based control group, likely due to the sealed, dry, nitrogen-filled environment and stable temperatures. This is a meaningful research outcome.

However, this was merely an experiment, not the starting point for a large-scale global rollout of underwater data centers. Microsoft did not commercialize Project Natick into a product line. Large tech companies typically do not leave clearly superior infrastructure options unused unless there are intractable real-world issues. These include maintainability, deployment logistics, submarine connectivity, permitting, environmental assessments, operational risks, and the fact that land-based data centers continue to improve.

China is moving faster than Microsoft in this field. An underwater data center project in Shanghai's Lingang area has entered commercial operation, reportedly targeting a capacity of 24 megawatts, combining seawater cooling with offshore wind power. Additionally, underwater data center development projects are underway around Hainan. These are real infrastructure projects, not laboratory demonstrations.

China has successfully built an underwater data center, proving the engineering feasibility of the technology. But this does not prove it will replace land-based data centers as the globally dominant architectural form.

Cooling is a major issue for data centers, especially as AI workloads push more power through denser chips. Underwater data centers can reduce mechanical cooling loads, improve power usage effectiveness, free up land, and alleviate pressure on local water systems. This is a valuable option for coastal locations that are hot, land-constrained, have strong marine engineering capabilities, and access to clean electricity nearby.

But data centers are about more than just cooling. They also involve power interconnection, fiber optics, land use, latency, tax bases, labor, water, construction, and increasingly, political factors. Moving racks into the ocean does not make the rest of these issues disappear; it only shifts the balance of constraints.

In a traditional data center, a failed server can be promptly replaced by a technician. A failed server inside a sealed module on the seabed is handled quite differently. Modules may need to be designed to operate without maintenance until planned retrieval, tolerate losses through redundancy, or be shipped back to shore by vessel. All these design choices involve costs, availability, and operational consequences.

Underwater data centers still require power and data connections. Submarine cables are mature infrastructure, but involve routing, landing points, permitting, repair vessels, security concerns, and failure modes. The complexity of marine infrastructure should not be underestimated.

Environmental issues require case-by-case analysis. Sealed underwater modules are not automatically an ecological disaster, nor are they automatically benign simply because "the ocean is big." Local impacts include seabed disturbance, thermal plumes, antifouling, noise, electromagnetic effects from cables, construction impacts, retrieval, corrosion protection, and more. Answers will vary by location.

Cybersecurity is also a dimension to consider. Researchers at the University of Florida and collaborators have shown that underwater data center systems may be vulnerable to acoustic attacks, as sound travels well in water and can potentially interfere with hard disk drives and system operations. This does not mean every underwater data center will crash from a single speaker, but it does introduce a different attack surface.

The current AI data center boom is putting every fringe infrastructure idea in the spotlight. Floating wave-energy AI data centers are attracting funding and headlines. Underwater pods are being repackaged as climate infrastructure. Offshore wind, seawater cooling, sovereign computing, edge latency, and AI demand are converging into a compelling narrative.

The real question is not whether underwater data centers can work, but whether they are better than land-based alternatives at specific locations, considering power supply, interconnection, cooling, land, permitting, fiber optics, maintenance, utilization, financing, environmental assessments, security, and lifecycle replacement.

For most of the world, the answer is likely no for some time. Land-based data centers can use efficient air cooling, liquid cooling, immersion cooling, waste heat recovery, and are closer to clean power and transmission capacity. They can be built in phases, continuously maintained, financed conventionally, and expanded without chartering ocean vessels. These numerous conventional advantages are highly competitive in infrastructure trade-offs.

For certain locations, underwater data centers may make sense. Dense coastal cities with expensive land, limited cooling water, strong port infrastructure, nearby clean electricity, and policy willingness to develop submarine industrial capabilities could justify them. China's coastal industrial zones are reasonable candidates for this reason. Some island grids or specialized military, telecommunications, or edge computing domains may also find value.

Underwater data centers may represent a real niche market, not a global reset of digital infrastructure. Cooling advantages alone are insufficient to make them a mainstream choice.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com