Battery Price Break-Even (Sezibwa)

[!info] Conservative basis. Tariff: UEDCL Code 10.2 Q3 2026 (666.5 / 562.1 / 429.7, configurable). Turnkey capex, real weather. Independent of the other docs because it rests on a battery-price assumption.

The question

At today's battery price, adding battery capacity lowers the return: the money is better spent on solar. Aaron's point stands: the long-term microgrid model wants more storage, so the real question is a price one. How cheap does battery have to get before expanding it pays? That turns the result into a procurement target.

Base case is no economically usable export for Sezibwa (ERA's Net Metering Regulations 2024 exist, but eligibility, interconnection and credit value are unresolved and typically pay below retail). So this asks only what storage is worth for shifting onsite solar into the evening.

Method

The same 20-year model (real load, real weather, battery aging to 50 percent by year 20, turnkey solar, Q3 2026 tariff, 15 percent discount) with the battery price swept as a variable. At each price the model re-finds the best allocation and compares its NPV against the best allocation that adds no battery. Today's price: SR-SE10B is 5,568,000 UGX for 8.9 kWh usable, about 626,000 UGX per kWh.

Result

The break-even is anchor-dependent (it depends how much solar you already have, because a bigger array spills more surplus a battery could soak up). On the current tariff:

At the 5 kWp optimum, battery must fall to about 367,000 UGX/kWh (roughly 40 percent cheaper) before the first extra unit pays. (On the compressed Q3 tariff this threshold is lower than an earlier draft's estimate, because storage now shifts cheaper energy.)

Investment metrics (at cheaper battery)

Findings

1. Break-even at the 5 kWp optimum is about 367,000 UGX/kWh on the current tariff — higher if you have already scaled to 6 kWp, lower at 4 kWp.
2. For material value you want battery well below break-even, toward ~200,000 UGX/kWh.
3. What this validates is "cheap storage," not "second-life batteries." Very low-cost storage would make a storage-heavy configuration attractive. But this model does not yet validate recycled/EV cells as the route there: a second-life battery starts already partially aged, so the fade-to-50-percent-over-20-years assumption is optimistic for it. Whether recycled cells actually deliver low installed cost needs its own model (acquisition, testing and grading, BMS integration, pack assembly, safety, warranty risk, heterogeneous cell health, and a shorter remaining life). Promising avenue, but the finance should not run ahead of that evidence.

Toward a break-even surface

A single break-even price is an oversimplification. The right output is a surface over four axes: battery price x remaining life x solar capacity x reliability/outage value. Remaining life matters as much as price: for example, 250,000 UGX/kWh for a cell expected to last 5 years can be worse than 400,000 UGX/kWh for one expected to last 12 years. And once the battery's backup value in outages is monetised, the threshold price rises. This surface, computed per site, is what NFE's site optimizer should eventually output.

Caveats

This prices energy value only; battery also buys backup, not counted here. Second-life economics need the separate model above before recycled storage is treated as validated.

Related

Builds on Optimal Solar and Battery Allocation. Companion cases: Battery Arbitrage, Solar Plus Battery.