
Source : https://www.networkocean.io/
As data demand (especially for AI) surges, traditional on-land data centers face increasing challenges: high cooling costs, land and real estate constraints, and massive water consumption. NetworkOcean’s solution taps into the ocean’s thermal properties. Their underwater capsules and floating barges use seawater cooling to deliver more efficient operations which they claim have shown up to 40% lower power usage. They aim not just to be green, but also more cost-efficient and fast to deploy, potentially unlocking high-density compute at lower operating costs.
Founded by Sam Mendel & Eric Kim, NetworkOcean is a hard-tech / climate-tech startup building and operating underwater data centers, with a focus on sustainability and efficiency. By submerging data centers in seawater, they leverage natural cooling to drastically reduce power consumption and eliminate freshwater usage. Their infrastructure is designed for high-density GPU workloads, making it particularly relevant for AI compute.

Source : https://www.networkocean.io/
NetworkOcean develops underwater data centre infrastructure and floating data centres specifically designed for high-performance computing needs, particularly GPUs for AI. They build watertight data centre capsules and data centre barges that utilise the natural temperature of the sea as a passive cooling system. Each unit is designed to accommodate high-capacity GPU racks, has a lower hardware failure rate due to the stable environment, and can be deployed quickly in about four months. With energy efficiency, low operating costs, and the ability to provide low latency for users in coastal areas, NetworkOcean’s products offer a new alternative that is more environmentally friendly and economical than traditional land-based data centres.
Looking ahead, NetworkOcean is preparing for a test deployment in San Francisco Bay, aiming to submerge their 500 kW capsule to validate performance, cooling efficiency, and environmental impact. The company is also beginning to offer GPU rack rentals (e.g., Nvidia H100) for AI clients, hinting at early commercial traction.
This combination of Y Combinator backing, a lean founding team, early funding, and a novel, sustainability-oriented infrastructure model positions NetworkOcean as a promising player in the emerging space of underwater / ocean-cooled data centers, offering a compelling alternative to traditional, resource-intensive data infrastructure.
Final thought
NetworkOcean stands at the intersection of climate-tech innovation and the escalating global demand for AI compute. By combining ocean-cooled infrastructure with modular, rapidly deployable data center capsules, the company demonstrates a bold rethinking of how digital infrastructure can scale sustainably. Backed by Y Combinator and supported by a forward-leaning engineering vision, NetworkOcean is not only addressing the energy and environmental challenges of traditional data centers, but also shaping a new blueprint for high-density, low-impact AI infrastructure. As AI workloads continue to surge worldwide, NetworkOcean’s approach signals a promising path toward greener, more efficient, and future-ready data computation.
References
Fondo. (2024). NetworkOcean launches underwater data centers. Retrieved from https://www.fondo.com/blog/networkocean-launches
Read Unicorner. (2024). NetworkOcean — YC S24. Retrieved from https://read.unicorner.news/p/networkocean
Startup Seeker. (2024). NetworkOcean company profile. Retrieved from https://startup-seeker.com/company/networkocean~io
VCBacked. (2024). NetworkOcean funding overview. Retrieved from https://www.vcbacked.co/company/networkocean
Y Combinator. (2024). NetworkOcean company page. Retrieved from https://www.ycombinator.com/companies/networkocean
NetworkOcean. (2024). Official website. Retrieved from https://www.networkocean.io
Crunchbase. (2024). NetworkOcean – funding and company overview. Retrieved from https://www.crunchbase.com/organization/networkocean
Contributor : Dyah Ayuning Tyas
Reviewer : Imam Buchari




Leave a comment