How to size a solar system under net billing — and why bigger stopped being better

Until February 2026, the sensible advice was to fill the roof. Under net metering, a unit you exported offset a unit you imported at close to parity — the meter effectively ran backwards, so surplus generation was stored on the grid at full value. Sizing generously was rational.

Net billing ended that. Export and import are now priced separately, and exported units are bought back at roughly a third of the old rate. The unit you consume as you generate it is still worth full retail tariff. The unit you push onto the grid is worth substantially less. Those are now two different products, and a system should be sized for the first one.

The arithmetic that changed

Take a site drawing 100 units between 9am and 4pm. Under the old regime, a system generating 160 units across the day banked the 60-unit surplus at near parity — the extra capacity paid for itself at close to the same rate as the first 100 units did. Under net billing that surplus earns about a third as much, so the marginal panels take roughly three times as long to pay back as the ones covering your own load.

The system does not become a bad investment. The last increment of it does. That is a sizing problem, not a solar problem, and the fix is to know your daytime consumption before anyone quotes you a kilowatt figure.

Which sites still pay back quickly

The businesses that suit solar best after net billing are the ones whose load curve already matches the generation curve. Their consumption peaks when the sun does, so very little is exported and the buyback rate barely touches them.

• Flour mills and rice mills — grinding runs through the working day.
• Cold storage and ice plants — compressor duty rises with ambient temperature, so the load peaks on exactly the hot, bright afternoons when the array produces most.
• Manufacturing on a single day shift — the plant and the array run the same hours.
• Offices and retail — air conditioning load tracks daylight closely.

The clearest example I have delivered is a cold store: see the 143 kW system at Ice Cold Storage, DG Khan, where compressor duty rises with the afternoon heat and the consumption curve tracks the generation curve almost exactly.

The harder cases are sites whose demand is mostly in the evening — some hotels, restaurants, and households where everyone is out until 6pm. Solar can still work there, but the case rests on either shifting load into daylight or on storage, and both need to be argued with numbers rather than assumed.

Does a battery fix the export problem?

Sometimes, and less often than you will be told. The pitch is simple: instead of exporting a unit at the buyback rate, store it and use it in the evening at the full retail rate. The gap between those two numbers is the arbitrage.

The problem is that the gap has to cover the cost of the storage over its cycle life, and on current LiFePO4 prices it frequently does not. Where a battery does earn its place is where there is a genuine outage cost — a cold store that spoils stock during load shedding, a process that cannot be interrupted safely. That is a reliability case, and it is a real one. Arbitrage alone usually is not.

What to have ready before you ask anyone for a design

1. Twelve months of electricity bills — not one. Seasonality is most of the picture.
2. Your sanctioned load and phase, from the bill or the connection agreement.
3. A rough sense of which hours your heavy equipment actually runs.
4. Your roof or ground area, and whether anything shades it during the middle of the day.

With those four things a competent engineer can tell you the right size within a fairly narrow band. Without them, any kilowatt figure you are given is a guess dressed as a quotation.