Coatue enters AI data-center land grab
- Coatue has launched Next Frontier, a new venture to buy land near major power sources and turn it into AI data-center capacity. - The key tie-up is with Fluidstack, the infrastructure company behind Anthropic’s $50 billion U.S. data-center buildout announced in November 2025. - This matters because AI’s bottleneck is shifting from chips to power, land, and grid access — and finance firms now want in.
AI data centers are turning into a land-and-power race, not just a chip race. That is the real story behind Coatue’s new move. The investment firm has launched a venture called Next Frontier to buy land near large power sources and turn those sites into future AI data-center campuses, with Fluidstack already involved as a partner and Anthropic seen as a likely customer. (techcrunch.com) ### Why is a hedge fund buying land? Because the scarce thing in AI is no longer just Nvidia GPUs. It is the whole physical stack around them — land that can support a huge campus, power that can be delivered on time, transmission access, cooling, permits, and construction capacity. If you thin(techcrunch.com 1)(techcrunch.com 2) ### What exactly did Coatue set up? Coatue created Next Frontier as a separate vehicle focused on acquiring parcels near major power sources, then developing or preparing them for AI data centers. The firm has already signed a joint venture with Fluidstack. That matters because Fluidstack is not just a random operator — it already has one of the biggest named AI buildouts on the board. (techcrunch.com) ### Why does Fluidstack matter so much? Fluidstack is the company Anthropic picked for a $50 billion U.S. infrastructure push announced on November 12, 2025. The first sites are in Texas and New York, with more to come, and the facilities are meant to be custom-built for Anthropic’s workloads. (techcrunch.com) labs will need more dedicated capacity than the cloud market can easily provide. (anthropic.com) ### Why isn’t the cloud enough? Because the biggest AI customers increasingly want bespoke capacity. Training frontier models and serving them at scale can require dedicated clusters, custom network layouts, long-term power planning, and tighter control over cost and uptime. A generic lease in a regular colocatio(anthropic.com)and more like private industrial plants for compute. (anthropic.com) ### What is Coatue really betting on? Coatue already has exposure to the AI boom through stakes in companies like Anthropic, OpenAI, xAI, DayOne, and CoreWeave. Next Frontier extends that thesis one layer down the stack. Instead of only betting on model labs and cloud providers, Coatue is betting that the hard as(anthropic.com)n profile, but it is the same AI conviction. (techcrunch.com) ### Why does this change the market? Because it brings more private capital into a bottleneck market. Hyperscalers were already competing for substations, utility commitments, and skilled builders. Now financial firms want those same choke points. That can speed projects up if capital helps ass(techcrunch.com)not pre-buy years of power and construction capacity. This is the land grab part. (techcrunch.com) ### What is the bottom line? Coatue is not just funding AI anymore. It is trying to own a slice of the physical bottleneck. That is a sign the market has matured into something more concrete and more constrained. In the next phase of AI, the winners may not just be the labs with the best models(techcrunch.com)egawatts first. (techcrunch.com)
