
Hi Y’all,
We’re Sam, Bingbin and James and we’re building AI datacenters out on the ocean, powered by the wind and cooled by the sea. Our first unit, A1DC, will hit the water by December and we already have customers lined up to rent the GPUs.
TLDR:
Ask:
Can you introduce us to neoclouds, AI labs, inference providers or other compute aggregators who would be interested buying compute on our first 10MW unit? Point them to: aikidotechnologies.com/reserve
Problem:
Solution:
Moving the units out into the ocean solves multiple challenges simultaneously:
Product:
Aikido’s product is AO60DC, which we build in farms of 10s-100s. Each platform is the size of a football field and supports:
Timeline:
We started designing our first unit, A1DC, back in February; kicked off retrofit of the offshore platform in the US in June, received final permits for the platform in September and will be doing a test fit-up in November.
We’ve already completed the first stage of the full-scale AO60DC design and will be kicking off the final design, with third-party certification, later this year. We will begin construction next Spring and the unit will hit the water late next year as we look for first customers to rent out the capacity.
The Team’s Backstory:
Bingbin and Sam met back in 2014 at a company that helped build some of the world’s first floating offshore wind, with now 100MW+ in operation. They decided to leave to make offshore wind platforms that were better/cheaper/faster than conventional ones, which were just borrowed from O&G. So they left and joined Otherlab, where they met James, who brings a background in aerospace, designing and building helicopters. They founded Aikido in 2022 with the support of Breakthrough Energy Fellows program (happy to refer folks!) and pivoted from the energy space to the datacenter space in early 2025.
The three of us with our first 1:250 scale prototype at a friend’s pool in 2022.
Clean marine AI datacenters
We build and deploy datacenters, out in the ocean, powered freely by the wind and cooled by the waves. We can serve both training and inference markets