In late February, when a massive storm brought more than a hundred inches of snow to the Sierra Nevadas, closing highways and paralyzing regional infrastructure, the product development team at Redwood Materials was closely monitoring the performance of their 12-megawatt pilot near Reno. It was a test of how their novel approach to powering a data center would hold up against severe weather.
Nancy Sun, the engineering VP leading the buildout of Redwood’s energy storage offering, wasn’t worried: In designing the pilot, her team had used 25 years of historical weather data at the site to model everything from ambient temperature to solar radiation.
That said, it was a live validation exercise for the site, which used a 20 MW solar array and repurposed EV batteries to power four modular Crusoe data centers.
Ultimately, all went smoothly, which Sun said the team took as validation of their design mantra, “ruthlessly simplify.” The system didn’t lose uptime or require heavy equipment to clear the way for maintenance. The site’s performance over those few February days is now feeding into the battery algorithms, sizing assumptions, and reliability models for future projects, Sun explained.
And last month, Crusoe opted to expand its pilot site to 24 data centers.
Founded by former Tesla executive JB Straubel, Redwood Materials was for years one of the country’s most high-profile battery recycling companies. Back when Sun joined Redwood in early 2024, an energy storage product wasn’t really on the table. Sun — a power electronics and hardware systems expert who previously founded and sold an autonomous trucking startup — was brought on to explore whether the company could or should expand from materials processing into product development.
Within about six months, she explained, energy storage had emerged as the natural product direction, leveraging existing battery inflows, testing and materials expertise, and operational footprint. Last summer it publicly pivoted into grid storage unveiling Redwood Energy at the same time as the Crusoe pilot announcement.
The opportunity, Redwood believes, is massive: According to the company, by 2030, end-of-life batteries could supply more than 50% of the entire energy storage market. But what, practically speaking, is involved when a company that has for years operated in recycling rather than storage builds a new product to take advantage of the moment?
Designing for simplicity
Redwood’s progression from a recycler to a BESS developer happened rapidly — over the course of about six months from idea to deployment, Sun said. Now, with the pilot up and running and delivering operational data, the question becomes how to scale it.
The unlock to Redwood’s grid storage setup is its proprietary “pack manager” technology, which lets the company communicate with and control batteries of different voltages, protocols and states of health, essentially making its grid offering battery-agnostic. But that’s not where Sun and her team started the development process, in part because they didn’t realize they’d need to build that software up front.
Instead, they started with the battery pack, wiring into real second-life packs that came in for processing, and experimenting with how to have them push and pull power to and from the grid. It wasn’t until they built a small pilot at one of Redwood’s facilities that they realized existing hardware and software didn’t meet their specific needs. To use second-life batteries, they’d need to build their own power electronics, their own controls, and software.
That’s how the core pack manager and site controller set was born. The pack manager is essentially a universal adapter for EV batteries, performing power conversion into a unified direct current stream that can pair with any standard inverter. Meanwhile, the site controller aggregates many packs into one coherent storage system, balances state of charge across packs, and intentionally “unbalances” based on health.
Older packs are cycled more gently, while fresher packs get worked harder, Sun explained. The battery algorithms team is continually updating models to track the health of each pack from receipt through its repurposed life — and, eventually, back to recycling.
Behind the product design team’s mantra of “ruthlessly simplify” is a focus on keeping both deployment and maintenance costs low, Sun explained. Traditional energy storage systems are high density and HVAC-heavy, adding cost and maintenance headaches. To avoid this, Redwood’s team opted for an open-air, low-density system mounted on above-ground cable trays.
Spreading packs out in the open air helps avoid the need for active refrigeration, and stripping away moving parts like fans and filters minimizes potential reliability failures. Keeping wiring above ground and limiting the size of each modular component minimizes the need for large equipment. As Sun explained, the result is a storage system that’s faster to build, easier to inspect after storms, and cheaper to keep running over time.
One key product development decision was to run the second-life battery packs at only a fraction of their original power capability; it was a trade-off, Sun said, between maximum output versus safety, longevity, and easier thermal management. This ended up being key to avoiding having to run an active refrigeration system, even in hot Nevada summers.
Building Redwood Energy
The original Redwood Materials business was geared toward massive machines and chemical processing capabilities, and a grid storage offering required a shift toward high-precision hardware development and embedded engineering, and microsecond-level power electronics controls.
But many of the key functions for a development team already existed within Redwood’s recycling and materials backbone, Sun explained, meaning she didn’t have to build the product development team from scratch. Operations staff at existing plants were cross-trained for power electronics manufacturing, and the company’s proprietary asset-tracking software was adapted to monitor the health of second-life batteries from receipt to deployment.
They did need to hire new domain specialists for the bespoke hardware. Sun filled those gaps with experts in power electronics and embedded software, which were specialities absent from Redwood’s chemical processing roots.
Today, Sun’s team is scaling production and designing for the “huge pipeline of projects” they have planned, as data center developers and utilities alike look for faster, lower-cost routes to power. Those include deployments that are more than 10 times the size of the newly-expanded Crusoe pilot.
The company has more than a gigawatt-hour of grid-ready batteries ready to go, and expects to add another five this year. Ultimately, the bottleneck to scale isn’t battery supply — Redwood receives over 20 GWh of batteries annually — but rather the safety qualification process required to put batteries out in the field. Sun explained that the core architecture, racking, pack manager, and software can handle multiple chemistries and form factors, but getting them past hurdles like the rigorous UL 9540A fire safety testing Redwood recently passed in-house is another matter.
Each type of battery has to go through a rigorous testing process. Qualifying each is time and resource intensive, so to start with, the team focused on a handful of pack types where they had the most inventory.
Over time, Sun said, Redwood is working to build what she calls a “qualification flywheel,” where the hard‑won processes, test infrastructure, and in‑house safety capabilities make it faster to bring new EV packs into the system — and so the limiting factor becomes how many batteries the world can send them, not how quickly Redwood can put them back on the grid.
