A company called Subsea Cloud plans to have a commercially available underwater data center off the coast of the United States before the end of 2022, with further deployments planned for the Gulf of Mexico and the North Sea. .
Subsea, which says it has already deployed its technology with “a friendly government faction”, plans to launch its first commercial pod before the end of this year near Port Angeles, Washington.
The company says placing its data center modules underwater can reduce power consumption and carbon dioxide emissions by 40%, as well as reduce latency by allowing the data center to be located closer. metropolitan areas, many of which are located near the coast.
However, according to Subsea founder Maxie Reynolds, it can also deploy 1MW of capacity for up to 90% less cost than it would take to commission 1MW in an onshore facility.
An illustration showing one of his commercial pods. Photo courtesy of Subsea
“The savings are the result of a smaller hardware BOM and lower deployment and maintenance complexities,” Reynolds told us. “It’s complex and expensive to set up in metropolitan areas, and in rural areas too: there are land rights and permits to consider and labor is slower and can be more expensive. “
The Port Angeles deployment, known as Jules Verne, will include a 20ft pod, similar in size and dimensions to a standard 20ft shipping container (one TEU or twenty foot equivalent) . Inside, there’s room for about 16 data center racks that can accommodate about 800 servers, according to Subsea. Additional capacity, if needed, is provided by adding another pod. The pod-to-shore link in this deployment provides a 100 Gbps connection.
As this is a commercial rollout, Jules Verne will be open to any potential customers or partners to come and check it out, virtually or not, according to Reynolds. It will be located in shallow water, visible from the port, while the Njord01 pod in the Gulf of Mexico and the Manannan pod in the North Sea are expected to be deeper at 700-900ft and 600-700ft respectively.
However, Jules Verne is unlikely to be used by many customers, as Subsea expects to use it primarily to demonstrate compliance to organizations and advocates who will inspect the pod and site, which could disrupt customer operations.
“We’re in talks with two of the well-known hyperscalers, so it’s still a bit up in the air,” Reynolds said.
Underwater pods are kept cool by being submerged in water, which is one of the reasons for reduced power and CO2 emissions. Inside, the servers are also immersed in a dielectric coolant, which conducts heat but not electricity. However, underwater modules are designed to passively disperse heat, rather than using pumps as is typically the case with immersion cooling in land-based data centers.
But what if there is a problem or if a customer wants to replace their servers? According to Subsea, customers can schedule periodic maintenance, including server replacement, and the company says it would take 4 to 16 hours for a team to travel to the site, set up the required pod(s) and replace. any equipment.
The viability of underwater data centers has already been demonstrated by Microsoft, which has deployed several over the past decade as part of its Project Natick experiment. The most recent was recovered from the seabed off the Scottish Orkney Islands in 2020 and contained 12 racks with 864 servers. Unlike the underwater modules, the Natick Project enclosure was filled with nitrogen.
Microsoft reported that only a “handful” of servers failed in its experience, and Subsea expects its data centers to require less maintenance due to reduced risk of environmental contamination like dust and dust. debris and reducing thermal shock.
Subsea said it plans to collocate data centers at sites offering various types of renewable energy infrastructure, and aims to have its data centers consuming only renewable energy by 2026. ®
Commercial Underwater Data Center Goes Online This Year