
Stop a battery failure before it becomes a fire
STRATA (System for Tracking, Routing, and Autonomous Task Administration) is being designed to independently supervise battery energy storage from any vendor, predicting thermal runaway and isolating failing racks before a local fault can spread through the site.
Battery storage scaled; its safety layer didn’t
AI data centers are installing battery storage behind the meter, next to billions of dollars of computing equipment. The battery management systems inside each unit run much the same control loop whether the site is 2 MWh or 400 MWh.
Those systems have three limits:
- Reactive: they alert after a thermal event has begun.
- Vendor-siloed: they can’t see across a site that mixes equipment from several manufacturers.
- Context-blind: they manage the battery box, not the facility around it.
When a fire does happen, the cost goes far past the equipment: downtime for the computing it powers, contamination, and local bans on new storage.
What STRATA is designed to do
- Fuses outside sensors (off-gas, infrared, ambient temperature, HVAC) with the battery system’s own data
- Uses AI to spot the early anomalies that come before thermal runaway
- Triggers an independent, hardwired emergency stop to isolate a failing rack
- Gives operators, insurers and fire authorities one independent record
Who it’s for
| Audience | What it is designed to give them |
|---|---|
| Data center and facility operators | Uptime protection for the computing load the batteries support |
| Integrators and manufacturers | An independent safety layer that adds to their own software rather than competing with it, or a white-label version under their own brand |
| Insurers | Independent evidence of risk controls before and after underwriting |
| Defense installations | Energy resilience: independent supervision of on-base battery storage and microgrids |
| Fire authorities | A verifiable safety record for permitting |
Why now
The build-out is already here: the U.S. installed a record 20.2 GWh of energy storage in the second quarter of 2026, and 30.8 GWh in the first half (SEIA and Benchmark Mineral Intelligence). Safety supervision has not kept pace.
Battery storage demand is expected to roughly double by 2030, from 271 GWh to 560 GWh, with AI data centers among the fastest-growing buyers. Insurers are already paying third parties to watch battery risk independently.
Operating battery storage at a data center, utility or installation?
STRATA’s core components have been demonstrated on synthesized data (technology readiness level 5). We are selecting sandbox partners to validate battery storage supervision on real equipment, and to shape it along the way.