Subsea Cloud: Forging a new path
Maxie Reynolds speaks to Natalie Bannerman about founding Subsea Cloud and how her company will disrupt the data centre space
Placing data centres underwater seems to be an obvious idea, given the growing challenges around data centre cooling and concerns about the environment, but until recently they only existed in theory or as experimental concepts. Then there is Subsea Cloud, a US-based company that is set to submerge its first commercial underwater data centre by the end of the year. Maxie Reynolds, Subsea Cloud’s founder and chief executive, explains how this innovative, albeit obvious, idea came true.
Reynolds’ started her career in a technical position for a company that built subsea tooling equipment, such as remotely operated vehicles and protection grounding units, for the offshore industry. From there she moved into the world of ethical hacking, before entering a subset of cybersecurity called red teaming. This work saw Reynolds and her colleagues test organisations’ cybersecurity by carrying out cyberattacks at them.
“One of the things that was coming up constantly were data centres, and I thought there aren’t that many things that you could put under the sea, but data centres are one of them,” says Reynolds.
After having that idea, she came up a design for a subsea data centre, and sent it to the very first person she work for in the offshore industry. He tweaked Reynolds’ design, her concept was validated, and Subsea Cloud was born.
Benefits of water
The biggest advantages Subsea Cloud’s approach to data centres offers are efficiencies in power, security and costs.
“It costs us a lot less to put one megawatt down into the ocean – about $600,000. To put one megawatt up and running on land costs about $6 million,” says Reynolds.
This cost difference is down to the bill for materials. “We don’t have that many materials,” explains Reynolds. “There aren’t many that you’d want to put into the ocean that can withstand it and really be welcomed into the environment.”
Also, Subsea Cloud does not need to buy land, which is another cost saving, and, as a bonus, Reynolds says it means her company “circumvents those land scarcities” that affect terrestrial construction projects.
In addition, Subsea Cloud’s subsea data centre pods can be deployed in just 12 weeks. And Reynolds says that once a pod is submerged deep enough, its physical security becomes absolute as “you can’t even get divers into those depths”.
When it comes to sustainability, Reynolds says Subsea Cloud’s data centres “are neutral to net positive for the environment” as the heat they generate does not interfere with the environment.
“We’re not going to heat the ocean and there’s science behind that,” says Reynolds. “It takes so much more energy to heat water by one degree than it does to heat the same mass of air.”
To use an analogy of Reynolds, if you jump into a swimming pool during a sweltering day the water feels freezing cold. This is because hot sunlight cannot heat a body of water to the same temperature as air.
“We eliminate 40% of electrically-driven cooling, and for each megawatt of power that is generated and used, half a tonne of carbon is created,” says Reynolds. “So, we’re stopping about 750 tonnes of carbon per pod, per year, from being emitted.”
For Reynolds, the most interesting aspect of this concept is that it is leveraging “age old” techniques and innovations.
“We’ve been using this technology for nearly as long as the offshore industry has been around,” she says. “There are computers in the ocean already. Every subsea wellhead needs to talk to something. We just extended a known technology we have. It’s a different use case.”
The biggest concerns around Subsea Cloud’s underwater data centre pods revolve around maintenance.
“We have some contingencies,” says Reynolds. “We can deploy redundant pods, as some cloud providers want the same on-land redundancy.”
Subsea Cloud is also prepared for outages. Each submerged pod also has a range of sensors – including pressure, fluid measurement, electrochemical, optical levels, liquid immersion temperature, and seismic activity to name few – that provide Subsea Cloud with clear view of what is happening to it. And Reynolds says the company keeps a ship on standby “within a 12-hour sailing radius” of a pod that can “go and pick up the pod and [perform] maintenance” if there is a problem.
Subsea Cloud’s first commercial deployment will see the “Jules Verne” pod placed on the seabed off Port Angeles, Washington. The company also has pods in the North Sea in Scottish and Norwegian waters, and a prototype in India.
Subsea Cloud’s data centre pods are currently custom-built models, as it is scaling up its footprint. But Reynolds says the company will have a factory ready in the US by 2025.
“What’s extremely important for us for the next 12 months is to build partnerships with wind, nuclear and cloud providers,” says Reynolds. “Those are really important and strategic for us.”