Exowatt, which aims to use concentrated solar power to generate 1¢ per kWh electricity for data centers, raised a $120M Series A from a16z, Sam Altman, and more
When Hannan Happi started thinking about how to solve the AI power crisis, he kept one figure in mind: one cent per kilowatt-hour.
Context & Ripple Effects
Exowatt’s financing puts a generation-side answer to data-center power constraints alongside a growing set of infrastructure plays focused on using available capacity more efficiently. Related coverage includes software aimed at curbing data-center energy demand and grid-to-GPU power systems for data centers.
The distinction matters: Exowatt is funding dedicated concentrated-solar supply rather than software or power-delivery optimization. Its one-cent-per-kWh target makes the economics of energy supply—not just access to compute—a central part of the data-center buildout.
First-order effects
- Exowatt gains $120M to advance its concentrated-solar approach for data centers, with backing from a16z, Sam Altman, and other investors.
- Data-center operators evaluating power-constrained expansion gain another prospective dedicated-power supplier, though the reported cost target remains an objective rather than an achieved market price.
Second-order effects
- Energy-management and power-architecture providers will face a clearer comparison: reduce and manage demand, as in Verse’s data-center energy-management push, or help customers secure new generation capacity.
- A credible low-cost generation path could make site selection and power procurement more consequential for data-center economics, while increasing pressure on competing power solutions to demonstrate delivered cost and reliability.
Third-order effects
- If dedicated generation can be deployed at the economics Exowatt targets, AI infrastructure investment may increasingly bundle compute facilities with long-duration power assets rather than treat electricity as an external utility input.
- The broader market could split between operators that secure power through owned or contracted generation and those that rely on grid capacity, making execution on construction, interconnection, and operations a key differentiator.
The trend: AI data-center infrastructure is becoming an energy-and-finance problem as much as a compute-hardware problem, spawning both new generation projects and tools to optimize scarce power.