From a lab bench in Pune to 74 cities.

We started with a test rig, a borrowed oscilloscope and a question: why do working sites still lose power when the grid is larger than it has ever been?

The first packs

The early work was pack-level. We built LFP packs for light electric vehicles and learned what Indian duty cycles do to a battery: heat soak, partial charging, rough handling and a charging window that is never quite the same twice.

That work produced the first version of our BMS firmware and the first telemetry pipeline. It also produced the rule we still design against: a pack you cannot see is a pack you cannot trust.

The first sites

Commercial sites came next. A QSR outlet, then a clinic, then a dark store. Each one had the same shape of problem: a genset covering gaps it was never efficient at covering, and a bill that did not reflect how unreliable the supply actually was.

FluidPod was built around that shape. One unit, one installation visit, a reserve that holds the critical loads first.

The platform

As the installed base grew, the interesting problem moved from cells to coordination. Hundreds of batteries, each with its own reserve, can behave like one dispatchable resource if the software is careful. FluidTwin became that layer: visibility, forecasting, reserve policy and dispatch.

Today

Storage deployed across homes, shops, clinics, warehouses and fleets in 74 cities. The same three things are still in-house: the pack, the firmware and the software. The map got bigger; the principle did not.