Home / Technology / AI Demands Fuel Redwood's Rapid Energy Storage Growth
AI Demands Fuel Redwood's Rapid Energy Storage Growth
20 Feb
Summary
- Redwood Materials' energy storage unit is rapidly expanding.
- AI data centers face grid connection delays of over five years.
- Google and Nvidia invested $425 million in Redwood's venture.

Redwood Materials, founded in 2017 by former Tesla CTO JB Straubel, has rapidly expanded its energy storage business. This unit, which launched in June 2025, has become the fastest-growing segment of the battery recycling and materials startup, largely due to the escalating demand from AI data centers.
The company's San Francisco R&D lab has quadrupled in size to 55,000 square feet and now employs nearly 100 people. This facility focuses on integrating hardware, software, and power electronics for energy storage systems essential for AI computing and large-scale industrial applications.
Redwood's growth is supported by a recent $425 million Series E funding round, with participation from new investor Google and existing backer Nvidia. This capital infusion is intended to scale the energy storage business.
The surge in AI has created an unprecedented demand for reliable electricity, leading to significant challenges for data center developers connecting to the grid. Many are facing estimated wait times of over five years for grid connections, a bottleneck Redwood's energy storage solutions aim to alleviate.
Initially focused on battery recycling and producing battery materials like cathodes, Redwood expanded into energy storage by launching Redwood Energy last summer. This venture utilizes thousands of EV batteries collected through its recycling business to power companies. One of its first customers is Crusoe, which operates a modular data center in Texas, powered by a Redwood system generating 12 MW of power and 63 MWh of capacity using retired EV batteries.
Redwood is preparing for larger deployments, with projects in the hundreds of megawatt hours and pipelines extending to multiple gigawatt hours, catering to hyperscalers that require massive power supplies.




