geothermal

Deep geothermal startups aim to deliver abundant clean power for data centers

Deep geothermal startups are betting they can deliver terawatts of clean electricity nearly anywhere, just as data centers' hunger for power grows insatiable.

3 min readTechCrunch
Deep geothermal startups aim to deliver abundant clean power for data centers

The geothermal startups chasing terawatts of power for data centers are onto something real, but the real story is not about drilling deeper, it is about who learns to scale first. These newcomers are betting they can bypass the established players and deliver clean, always-on electricity almost anywhere. That is an audacious claim, and it deserves scrutiny, not hype. What matters for our readers is whether this technology can actually keep pace with the energy demands that AI itself is creating.

The connection between AI and energy is becoming impossible to ignore. As A top AI researcher joins Spotify's Daniel Ek to build fusion power shows, some of the brightest minds are already chasing fusion as the ultimate solution. Yet fusion remains years away from commercial reality. Geothermal, by contrast, is proven technology that simply needs to be made more accessible and more flexible. That is where the startups see their opening. They are not trying to reinvent the physics, they are trying to reinvent the deployment model. If they succeed, data centers could be built near any geological formation with sufficient heat, rather than only in regions with existing hydrothermal resources. That would transform where and how the AI infrastructure of the next decade gets built.

Our readers should watch one specific tension: the same AI workloads that demand more compute also generate new tools for managing infrastructure. Google Donates AI Agent Tools to Open-Source Foundation for Secure Scaling reminds us that AI itself is being turned toward operational challenges like security and scaling. If geothermal startups can integrate these kinds of intelligent agents into their drilling and monitoring workflows, they could accelerate their learning curve dramatically. The question is whether they will move fast enough to matter. The incumbents have decades of geological data and established supply chains. The newcomers have agility and a willingness to fail fast. That trade-off is the one to track.

The concrete takeaway is this: watch the next two years of pilot projects. If a startup can demonstrate a repeatable, cost-competitive drilling method in a location outside the traditional geothermal belts, the calculus for data center siting shifts permanently. If they cannot, the fusion timeline becomes the one that matters. Either way, the era of assuming data centers will always run on fossil-backed grid power is ending. Our readers should start asking their energy providers what the plan is for 2030, not 2050.

From TechCrunch

New geothermal startups are betting that they can leapfrog today's leaders to deliver terawatts of electricity nearly anywhere.

Read the original at TechCrunch