
Hybrid storage sounds like a way to get the best of two chemistries. In practice it is a control problem wearing a hardware costume — and it only pays when the two duties are genuinely different.
Where the split makes sense
A site with a long, slow daily cycle and a handful of large power spikes. Sodium covers the energy (cheap per cycle), lithium or a supercapacitor covers the spike (cheap per kW). The controller routes the spike so the sodium bank never sees it.
This is real: peak-shaving plus arbitrage on a site with an intermittent high load, such as a crane or a press.
Where it does not
If both chemistries see the same load, you have bought two battery systems, two BMS sets and a controller to shuffle energy between them — with round-trip losses at every shuffle.
A single chemistry sized for the worst case is often cheaper. Run the arithmetic before being impressed by the diagram.
The control strategy is the product
Who decides which bank serves which load? What happens at the boundaries — when the sodium bank is full and the spike arrives? How is state of charge balanced between banks with different curves?
These are software questions with hardware consequences, and they are where hybrid projects stall.
How to evaluate one
Ask for the control logic in writing, the round-trip efficiency including the converter paths, and a one-year simulation against your load profile. If the proposal has none of those, it is a diagram, not a system.
Two sodium families, two jobs
A low-temperature 50 Ah cell is specified at 2.85 V while the standard product of the same format sits at 3.0 V — different cathode and electrolyte, different cold behaviour. That is the kind of split a hybrid design can exploit: one bank tuned for cold charge acceptance, one for cycle cost.
Where to go next
Questions we get asked about this
Is hybrid storage worth the complexity?
Only when the load has two clearly different duties — long energy cycling and short power spikes. Otherwise a single chemistry is cheaper.
How is the split controlled?
An EMS routes each load to the bank that serves it best. The control logic, not the hardware, decides whether the system works.
What efficiency should I expect?
Lower than a single bank, because every transfer through a converter loses energy. Ask for the round-trip figure including all paths.

