- Unicorner
- Posts
- NetworkOcean
NetworkOcean
Building and operating underwater data centers
🦄 Unicorner Startup of the Week:
NetworkOcean
✍️ Notes from the Editors
Today's article is brought to you by Kalshi. You can now bet on elections in America. Put your money where your mouth is and place your first trade today.
Over the last two weeks, we brought together hundreds of founders and investors for SF and LA Tech Week.
So, at our latest event, we asked dozens of attendees: Why Unicorner?
Beyond simply being a newsletter, Unicorner is here to connect the greatest people with the greatest people. May that be through real-life events, or digitally to our tens of thousands of weekly readers, our community aims to be a value-add to anyone it reaches.
In that spirit, if after reading this week’s article, your immediate thought is “Wow! I wish I could invest” — reply to this email and we’ll patch you through to the founders.
Building and operating underwater data centers
NetworkOcean addresses the rising demand for data centers while also considering the growing power requirements. It focuses on mitigating the potential impacts of increased power usage, including higher emissions and strain on the grid. Offering underwater data centers, its goal is to offer cheaper and more sustainable solutions to data centers, all by taking advantage of water cooling in its infrastructure. With a claimed 25% lower operating cost and reduced energy needed to cool the increasingly hotter data centers, NetworkOcean aims to bring a new age for data centers.
🔗 Check it out: networkocean.io
You can now bet on elections in America.
Whether you want to hedge candidate risk, bet on your preferred party, or simply think you know something the public doesn’t–Kalshi is the place to put your money where your mouth is.
As the only regulated American exchange, Kalshi is the only option if you’re looking to avoid the next FTX. Open positions earn 4.05% interest, and with up to $100 million liquidity, Kalshi can easily fulfill large trades and institutional demand.
If that isn’t enough: Kalshi has suspended all fees on election markets.
Yep. $100 million in liquidity. 4.05% interest. Zero fees.
Join us for this historic moment and place your first trade today.
This is sponsored content. For sponsorship inquiries, please reply to this email.
💰 Business Model
Akin to other data centers, NetworkOcean plans to sell its capacity to those who need data centers and GPUs for various use cases, including training, inference, fine-tuning, and experiments. It aims to charge approximately $2.10 per GPU per hour, which is lower than the standard industry rate.
📈 Traction and Fundraising
Raised $600,000 Pre-Seed from Y Combinator, Refactor Capital, Z Fellows, Pareto, Climate Capital, and BullMont Capital
Participated in Y Combinator S24 batch
Has $1.2 million committed for an upcoming $5 million seed round
Has signed LOIs with six AI companies
Has a 1-megawatt data center near completion with projections for $10 million annual recurring revenue
👫 Founders
Sam Mendel, CEO: Previously founder and CEO @ CreatorCrew, Founder and CEO @ CASHU, Co-Founder and Growth + Production @ Rant!, Student @ Babson College, Student @ Olin College of Engineering
Eric Kim, CTO: Previously Software Engineer Intern @ Palantir, Software Engineer Intern @ Merge, Engineering Physics + Computer Science @ Cornell
💼 Opportunities
🔮 Our Analysis
Recent news on how major players in AI and data, such as Microsoft, Amazon, and Google, have gone as far as to sign deals to use and reactivate nuclear power has brought to the forefront the ever-increasing need for energy and data centers. In addition, Nvidia’s meteoric rise to dominate the computing market has also brought to light the high demand for more powerful and efficient GPUs for AI training, modeling, and other computational use cases.
Historically, data centers have been air-cooled, with fans replacing hot air with cold air. As these centers became more advanced, temperature levels rose, resulting in a greater need for cooling. While water cooling isn’t new (it’s been around for nearly 50 years), it hasn’t been a practical choice, given the increased costs it presents versus air cooling. However, with the proliferation of AI use cases, water cooling has dramatically increased as the choice of cooling for these data centers.
Data centers have also been a concern for environmentalists and companies due to the tremendous amounts of energy they consume. Companies have been looking for different ways to accommodate this increased AI spending while being environmentally minded. Many different strategies are being used by companies to fuel this need for power, looking at a variety of energy sources to help gain an edge in the growing AI race.
NetworkOcean builds on the prior innovation from Microsoft’s experiments with underwater data centers and has found effective ways to commercialize the idea. The complexity of building an underwater data center in the ocean is offset by saving energy and operational costs, where it has pioneered certain processes to maximize its operational efficiency. It claims a lower failure rate and greater reliability than on-land data centers. In addition, by taking advantage of the open ocean waters, it has also forgone the complications of using freshwater on land for water cooling, adding to its savings and lower price points. Ignoring its competitive prowess, it is also providing a sustainable solution to a growing need. Having gone through Y Combinator’s most recent summer class, the team is well-equipped to lead the charge in an exponentially growing market. Given its strong economic and environmental advancements, we believe NetworkOcean can turn the tide in data servers, becoming the backbone of future computing infrastructure.
📚 Further reading
Microsoft finds underwater datacenters are reliable, practical and use energy sustainably [Microsoft]
Hungry for Energy, Amazon, Google and Microsoft Turn to Nuclear Power [The New York Times]
Power-Guzzling AI to Drive ESG Debt Sales, Morgan Stanley Says [Bloomberg]
Unlocking Savings: Why Buying NVIDIA H100 GPUs Beat AWS Rental Costs [TRG Datacenters]
Written by Kanay Jay Shah
Is this week's company a future unicorn? |
Introducing Roku Ads Manager: CTV Ads Made Simple
Reach viewers where they’re already engaged on the big screen
Run ads on top ad-supported streaming channels
Get started today for as little as $500
Sponsored content
If you enjoyed today's article, forward this email to a friend!
If you're just seeing this email now, subscribe here.
Made with 💜 by the Unicorner team 🦄