China is preparing to unveil a groundbreaking $223 million underwater data center located around six miles from the coast of Shanghai. The facility, scheduled to begin operations in September, will be capable of hosting up to 792 high-performance AI servers. By using seawater for natural cooling, the center is expected to reduce electricity consumption by about 30% compared to conventional land-based data centers. Additionally, the project is designed to operate on 97% wind-generated energy, positioning it among the most energy-efficient large-scale data infrastructure projects globally.
The cutting-edge submerged facility is designed to accelerate AI model training at an unprecedented pace. Reports suggest it could process massive AI models, such as GPT-3.5, within a single day—a task that typically requires significantly more time on traditional systems. This leap in processing speed could give China a major technological edge in the competitive global AI race, especially as demand for faster, more sustainable computing solutions continues to rise.
The concept of underwater data centers is not entirely new. Microsoft’s “Project Natick” successfully tested a similar offshore cooling system in 2018, but the initiative was eventually put on hold. China, however, is taking the idea to the next level, rapidly scaling it for large-scale commercial and AI applications. This move signals a shift towards unconventional infrastructure solutions aimed at meeting the growing energy and cooling demands of next-generation computing.
While the project promises significant efficiency and sustainability benefits, environmental concerns remain. Experts are still assessing the potential impacts of such facilities on marine ecosystems. Despite these questions, interest in the technology is spreading across Asia. Both South Korea and Japan have begun exploring offshore cooling methods to prepare for an AI-driven future, suggesting that underwater data centers could soon become a common feature in global tech infrastructure.
