China is in the process of implementing an innovative project involving an underwater commercial data center near Hainan Island in the South China Sea.
This initiative is a collaborative effort between the China Offshore Oil Engineering Co. (COOEC), the nation’s leading Engineering, Procurement, Construction, and Installation (EPCI) company, and Highlander, a private data center company.
While specific details about the computing hardware remain undisclosed, Highlander asserts that each underwater module is capable of processing over four million high-definition (HD) images within a mere 30 seconds.
Enclosed within watertight storage modules weighing a total of 1,300 tons, these computing systems are submerged approximately 115 feet (35 m) underwater, a process taking around three hours per module.
The initial module installation is currently underway, but Highlander’s ambitious plans include deploying 100 such modules, establishing a capacity for nearly six million computers concurrently.
The significant heat generated by this extensive computing power will be naturally dissipated by the surrounding seawater, resulting in an anticipated saving of 122 million kilowatt-hours of electricity that would otherwise be required for land-based cooling.
Expected to be operational by 2025, the facility will also conserve 732,000 square feet (68,000 square meters) of terrestrial land, repurposable for other uses, and save 105,000 tons of fresh water used in conventional cooling systems.
Despite a 25-year lifespan designed for the modules, uncertainties persist regarding the impact of corrosive seawater and underwater ecosystems on their construction. Highlander’s experience in such installations is relatively limited, with tests conducted in January 2021 in the Guangdong port of Zhuhai.
In a similar vein, Jules Verne Pod, a US commercial subsea data center, was also explored in 2022. In comparison to conventional land-based data centers, it is claimed to be more affordable, more dependable, and delivers faster connectivity.
