South Korean city bets on the ocean to solve its AI data center crisis — but this is only the beginning
THE SO WHAT
A $37M underwater, seawater-cooled data center is a concrete move toward treating AI compute as coastal industrial infrastructure, not just cloud real estate. If your roadmap assumes unconstrained, cheap data center capacity on land, you need a parallel plan that prices in power, water, and siting constraints.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechFloating 10 MW nuclear reactor proposed for California: barge-mounted SMR aims to power 15,000 homes
A 10 MW barge-mounted SMR proposal in Long Beach is another data point that power for compute is pushing nuclear into more modular, locationally flexible formats. If you’re planning multi-decade AI or industrial capacity, you now have to track nuclear permitting and SMR timelines alongside grid and renewables.
Deep & Emerging TechSources: some of Nvidia's top customers have been told that prices will jump 15%+ on systems, including Vera Rubin and Grace Blackwell, starting in early 2027
A 15%+ price step-up on next-gen systems like Vera Rubin and Grace Blackwell turns GPU procurement into a forward-hedging problem, not just a budget line. Lock in multi-year terms where you can and get serious about model efficiency — FLOPs per dollar is about to matter as much as raw capability.
Deep & Emerging TechChina’s AI and chip firms are handing out shares to keep their engineers
When Cambricon is giving equity to 85.3% of staff and AMEC to more than 97%, engineering talent has effectively become the scarce national resource in semis. If you’re competing for similar profiles—especially in Europe, where the gap is pegged at 65,000 workers—assume equity is now table stakes, not a perk.
Deep & Emerging TechThe company that just opened America’s newest battery plant says solid-state cars are a decade away
If LG Energy Solution expects solid‑state in phones roughly a decade before EVs, grid‑scale and vehicle roadmaps need to assume today’s lithium‑ion constraints for longer than the hype suggests. Auto and mobility operators should plan around incremental density and safety gains, not a near‑term chemistry step‑change.