Ask Runable forDesign-Driven General AI AgentTry Runable For Free
Runable
Back to Blog
Technology6 min read

China sinks 2,000 servers beneath the ocean as AI power demands push data centers into entirely unfamiliar territory | TechRadar

China places a massive AI data center underwater, effectively reducing its PUE Discover insights about china sinks 2,000 servers beneath the ocean as ai power d

TechnologyInnovationBest PracticesGuideTutorial
China sinks 2,000 servers beneath the ocean as AI power demands push data centers into entirely unfamiliar territory | TechRadar
Listen to Article
0:00
0:00
0:00

China sinks 2,000 servers beneath the ocean as AI power demands push data centers into entirely unfamiliar territory | Tech Radar

Overview

News, deals, reviews, guides and more on the newest computing gadgets

Start exploring exclusive deals, expert advice and more

Details

Unlock and manage exclusive Techradar member rewards.

Unlock instant access to exclusive member features.

Get full access to premium articles, exclusive features and a growing list of member rewards.

China unveils 'world's first' underwater data center — 2,000 server facility is powered by offshore wind, and cooled by the sea, making it one of the most efficient around

China places a massive AI data center underwater, effectively reducing its PUE

When you purchase through links on our site, we may earn an affiliate commission. Here’s how it works.

China submerged nearly 2,000 AI servers beneath the ocean near Shanghai

Seawater now cools Chinese AI servers without traditional industrial chillers operating continuously

China connected offshore wind farms directly to an underwater artificial intelligence facility

China has begun commercial operations at an underwater data center where sealed server modules operate beneath the ocean using seawater for passive cooling.

The project combines offshore wind generation with subsea computing infrastructure to reduce electricity pressures linked to artificial intelligence expansion worldwide.

This underwater data center sits roughly 35 meters below the ocean surface near Shanghai’s Lingang Special Area and houses nearly 2,000 servers, including GPU clusters from China Telecom and Link Wise.

'We have this power from the wind. We have free cooling': This startup wants to build underwater data centers inside wind turbines at sea - using the icy North Sea waters to keep everything cool

How wave-powered ocean platforms could meet AI data center energy demands

Samsung’s floating data center plan promises faster AI power deployment at sea

Chinese authorities and private engineering company Hi Cloud Technology jointly developed the $226 million installation.

This 24-megawatt installation processes artificial intelligence workloads, 5G services, and large-scale data annotation operations requiring substantial computing capacity.

Unlike conventional land-based facilities using industrial cooling systems, the underwater structure depends heavily on naturally stable ocean temperatures surrounding pressure-resistant server modules.

Cooling demands have increasingly become a major obstacle for modern data centers because advanced GPU clusters generate enormous heat during continuous computing operations.

According to Chinese media reports, the underwater installation achieved a Power Usage Effectiveness (PUE) rating below 1.15, lower than the industry average hovering around 1.5.

A lower PUE indicates that more electricity supports computing tasks directly instead of auxiliary systems such as cooling equipment, ventilation, and infrastructure maintenance.

Industry analysts have increasingly examined alternative cooling methods because expanding artificial intelligence infrastructure continues placing pressure on national power grids and electricity availability.

How wave-powered ocean platforms could meet AI data center energy demands

China testing truck-mounted nuclear reactor built to power AI data centers for decades

Hitachi and Mitsui OSK Lines to develop floating data centers in Japan

The Shanghai project also reflects China’s broader effort to integrate renewable energy generation directly into digital infrastructure.

Offshore wind farms connected to the underwater facility reportedly provide a substantial share of operational electricity, reducing their dependence on conventional grid-based energy supplies.

Authorities described the project as the “world’s first” offshore wind-powered underwater data center operating at commercial scale, although underwater computing experiments already existed elsewhere.

Microsoft previously tested submerged data center capsules through its Project Natick initiative, conducted near Scotland and California before discontinuing commercial development efforts.

Those earlier experiments nevertheless suggested underwater systems could experience lower hardware failure rates because sealed environments limited exposure to oxygen and temperature fluctuations.

However, large-scale underwater deployments continue facing significant engineering concerns involving corrosion, pressure sealing, subsea cable durability, and long-term hardware accessibility during emergencies.

Replacing malfunctioning equipment underwater remains considerably more complicated than conventional facilities, where technicians can physically inspect servers and infrastructure within minutes.

Operators therefore depend heavily on remote monitoring technologies, modular sealed systems, and redundant infrastructure intended to minimize direct maintenance requirements throughout operational lifespans.

Similar concepts continue to emerge globally as governments and technology companies examine unconventional approaches for handling artificial intelligence infrastructure demands without overwhelming terrestrial resources.

Recent reports detailed how startup Panthalassa, backed by Peter Thiel, is developing floating data centers using wave energy and ocean water cooling systems.

Although underwater facilities may reduce cooling energy consumption substantially, long-term operational reliability remains uncertain because large commercial deployments remain relatively uncommon worldwide.

Follow Tech Radar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.

Efosa has been writing about technology for over 7 years, initially driven by curiosity but now fueled by a strong passion for the field. He holds both a Master's and a Ph D in sciences, which provided him with a solid foundation in analytical thinking.

You must confirm your public display name before commenting

1 High-end hi-fi company YBA has made what it claims is the world's first "transportable" CD player for SACDs — it's got the audiophiles in the office very excited, but it's not exactly a handheld model to slip in your jacket pocket

2 Nvidia really doesn't seem to care about gaming GPUs anymore — the company won't even bother to break down graphics sales in its big investor reports

3'A true engineering feat': Silicon-based qubits have a clear advantage in race to million-qubit quantum computer thanks to tech that made Nvidia AI GPUs possible

4 The biggest surprise from Google I/O 2026? Poll reveals majority of Android fans are most excited for Android XR and Gemini Spark— but we’re still waiting to hear more about Googlebooks and the Google Smart speaker

5 Empowering Enterprise AI with Dell and NVIDIA Whitepaper

Tech Radar is part of Future US Inc, an international media group and leading digital publisher. Visit our corporate site.

© Future US, Inc. Full 7th Floor, 130 West 42nd Street, New York, NY 10036.

Key Takeaways

  • News, deals, reviews, guides and more on the newest computing gadgets
  • Start exploring exclusive deals, expert advice and more
  • Unlock and manage exclusive Techradar member rewards
  • Unlock instant access to exclusive member features
  • Get full access to premium articles, exclusive features and a growing list of member rewards

Cut Costs with Runable

Cost savings are based on average monthly price per user for each app.

Which apps do you use?

Apps to replace

ChatGPTChatGPT
$20 / month
LovableLovable
$25 / month
Gamma AIGamma AI
$25 / month
HiggsFieldHiggsField
$49 / month
Leonardo AILeonardo AI
$12 / month
TOTAL$131 / month

Runable price = $9 / month

Saves $122 / month

Runable can save upto $1464 per year compared to the non-enterprise price of your apps.