Microsoft waves goodbye to underwater data centers
Microsoft has stopped using its underwater data centers that were part of an experiment to study conditions around machinery.
What you need to know
- Microsoft no longer has any active underwater data centers.
- The company experimented with underwater data centers as part of its Project Natick.
- Underwater data centers saw one-eighth of the failure rate compared to on-land counterparts.
- Microsoft explained that it will use what it learned from the project in other cases.
With the heat dome causing temperatures to soar in many parts of the world, jumping into a frigid sea may sound refreshing. But if you do jump into cold waters, you won't have the company of underwater data centers from Microsoft. The tech giant began testing underwater data centers in the North Sea in 2018, but it no longer has any active underwater data centers.
The undersea data centers were part of Microsoft's Project Natick, which aimed to test the feasibility of placing data centers underwater. The experiment saw giant tubes placed 117 feet below the surface of the Scottish sea. The unique environment of the datacenter allowed Microsoft to test several factors, such as filling the data center with nitrogen rather than oxygen.
Humans and machines have very different needs when it comes to creating an environment for optimal performance. Oxygen is essential for people, but it can cause damage to computers. Temperature requirements also differ greatly between mankind and machines. Keeping the data centers separate from humans also prevented anyone from bumping into components.
Project Natick saw interesting results, including underwater data centers having one-eighth of the failure rate of data centers with the same components on land.
While the results seemed promising, Microsoft does not have any active underwater data centers in 2024. The tech giant shared few details on why it has stopped using underwater data centers. Project Natick was an experiment, so there's always a chance Microsoft just wanted to learn from the project and take lessons elsewhere.
"I'm not building subsea data centers anywhere in the world," said Noelle Walsh, Corporate Vice President of Microsoft's Cloud Operations + Innovation team to Data Center Dynamics.
"My team worked on it, and it worked. We learned a lot about operations below sea level and vibration and impacts on the server. So we'll apply those learnings to other cases."
Get the Windows Central Newsletter
All the latest news, reviews, and guides for Windows and Xbox diehards.
While Microsoft does not have any underwater data centers at the moment, the tech giant plans to use Project Natick as a research platform. Microsoft explained the future of the project in a statement to Data Center Dynamics:
"While we don’t currently have data centers in the water, we will continue to use Project Natick as a research platform to explore, test, and validate new concepts around data center reliability and sustainability, for example with liquid immersion."
There's a good chance that Microsoft will use findings from its Project Natick to improve data centers in other environments.
An ever-growing demand for data centers
Insights from Project Natick and other experiments will become increasingly important going forward. There is a massive demand for data centers on a global scale, and that's only set to go up over time.
Artificial intelligence is everywhere in 2024. From your smartphone to your computer, AI features are increasingly common. While some devices, such as Copilot+ PCs, perform AI tasks locally, much of the data processing for AI is done in the cloud, creating an ever-increasing demand for data centers.
Of course, AI is far from the only type of technology driving the demand for data centers higher. Cloud computing, using a smart home device, and even searching the web require data centers. With billions of devices connected to the web, the demand for data centers almost feels infinite, and that demand is going to continue to grow.
Data centers require energy to operate and require maintenance. Reducing the amount of energy needed by data centers could go a long way in managing energy grids and helping spread resources to other services and devices.
Sean Endicott is a tech journalist at Windows Central, specializing in Windows, Microsoft software, AI, and PCs. He's covered major launches, from Windows 10 and 11 to the rise of AI tools like ChatGPT. Sean's journey began with the Lumia 740, leading to strong ties with app developers. Outside writing, he coaches American football, utilizing Microsoft services to manage his team. He studied broadcast journalism at Nottingham Trent University and is active on X @SeanEndicott_ and Threads @sean_endicott_.
-
GraniteStateColin Why did they decide to remove them? That's why I opened the article, but no information on that. E.g., Was it too expensive to service them, even with the lower failure rate? Was it concern over heating the water and causing environmental issues? Was it always intended just as an experiment to gather data and never to be a long-term approach?Reply -
Sean Endicott
Unfortunately, Microsoft hasn't shared details as to why it stopped using data centers. I've reached out to them to see if there are more details available. So far, we just know that the company does not have any active underwater datacenters.GraniteStateColin said:Why did they decide to remove them? That's why I opened the article, but no information on that. E.g., Was it too expensive to service them, even with the lower failure rate? Was it concern over heating the water and causing environmental issues? Was it always intended just as an experiment to gather data and never to be a long-term approach? -
fjtorres5591
The lower failure rate almost certainly came from the nitrogen atmosphere. That is applicable everywhere. Another (not mentioned) advantage is the data center modules were prebuilt and dropped into place. They might switch to standardized factory built, sealed modules. Maybe container sized.GraniteStateColin said:Why did they decide to remove them? That's why I opened the article, but no information on that. E.g., Was it too expensive to service them, even with the lower failure rate? Was it concern over heating the water and causing environmental issues? Was it always intended just as an experiment to gather data and never to be a long-term approach?and
Experiments end.
Looking at the photo in the OP, the underwater environent sems pretty corrosive, what with the barnacles and rust. Savings in cooling might be outweighed by manufacturing cost and higher maintenance cost (having to bring it up for upgrading, for example).
Just speculation but whatever they might have learned, if useful, will be a trade secret. Why help Amazon cut *their* costs? 😎 -
GraniteStateColin
Thanks for the update!Sean Endicott said:Unfortunately, Microsoft hasn't shared details as to why it stopped using data centers. I've reached out to them to see if there are more details available. So far, we just know that the company does not have any active underwater datacenters. -
GraniteStateColin fjtorres5591 said:The lower failure rate almost certainly came from the nitrogen atmosphere. That is applicable everywhere. Another (not mentioned) advantage is the data center modules were prebuilt and dropped into place. They might switch to standardized factory built, sealed modules. Maybe container sized.
Experiments end.
Looking at the photo in the OP, the underwater environent sems pretty corrosive, what with the barnacles and rust. Savings in cooling might be outweighed by manufacturing cost and higher maintenance cost (having to bring it up for upgrading, for example).
Just speculation but whatever they might have learned, if useful, will be a trade secret. Why help Amazon cut *their* costs? 😎
Yeah, those are consistent with my assumptions too, and (also to Sean's point), I'm not surprised MS isn't sharing those details. But, w/o them, we're left with our own speculation. Understandable, but also less interesting. :)