In March 2024, a customer asked for an answer to the usual solar vs gas generator question. The rural grant deadline was four days away. I had one phone call, no load logger, and a proposal that was already late. So I made the mistake I now warn everyone about: I compared the kW rating of a gas generator with the kW rating of a solar inverter and told the owner that gas would be cheaper to install.
Then I reopened the load list. Then I re-read the SMA 7000 inverter data sheet. Then I called back and told the customer not to sign the estimated $9,400 generator agreement.
I have been handling solar-plus-storage orders for commercial and small industrial clients since 2017. I personally made and documented 14 significant mistakes, totaling roughly $29,000 in wasted budget. This near miss did not cost the full amount, but it is the reason my team now has a pre-bid checklist.
Solar vs gas generator: the wrong starting level
The classic comparison starts with an output number: 7 kW gas generator versus 7 kW inverter. I have done that. Product literature encourages it. In reality, the two machines behave differently when customers connect things.
At least, that has been my experience on sites with motor loads, which are most rural B2B sites.
What appeared on my load sheet
The customer ran a small packing and tool maintenance business. The critical loads looked easy: a well pump, two reach-in coolers, a network cabinet, LED lights, and a 7.4V battery charger for tool batteries. Running total was around 4.6 kW. A 7 kW generator seemed to have enough margin.
The real worksheet cared less about running watts than starting surges. The list included:
- 1.5 hp well pump: around 1.4 kW running, with a starting surge near 6.2 kW
- two reach-in coolers: 0.9 kW running, combined start kick around 2.8 kW
- network and security equipment: 0.45 kW constant
- LED lighting and small shop equipment: 0.8 kW
- 7.4V battery charger for tool batteries: 0.1 kW
The charger was not the problem. The problem was sequence. If the coolers and charger were already running and the well pump started, the momentary load was closer to 10.4 kW. A 7 kW generator could not handle that without tripping or browning out the site. The gas generator quote would have needed a larger unit, not a smarter inverter label.
When I re-quoted solar, my first design included an SMA 7000 inverter because the solar side was 7 kW. It was not wrong to choose that model. It was wrong to compare it to a gas generator as if 7000 watts was the only spec that mattered. The inverter data sheet tells me the maximum input and continuous AC output, not what happens if the customer starts two induction motors at the same time.
Why the generator quote felt easy
The easier the comparison feels, the more likely somebody skipped a step. A gas generator can be quoted from a spec sheet and a concrete pad. A solar system needs more decisions: PV, battery, transfer, load panel, permits, and grid rules. But extra steps are not a reason to push a customer into a fuel-dependent hunt whenever the power goes out.
What the generator quote did not include was fuel after hour one, or the gasoline run to a station, or oil checks, or annual exercise, or the fact that longer outages often happen during road closures. Those costs do not appear on the compare line. For a short emergency, they are acceptable. For a customer who could lose power for days, they become the whole decision.
This is why the phrase 'AI power inverter generator' makes me pause. The product category is real and useful; inverter generators do govern engine speed and clean the waveform. But the AI part is usually load management, not energy creation. No algorithm removes the need to store fuel. According to FTC business guidance (ftc.gov), product claims should be truthful and substantiated. I try to hold my proposals to that standard as well.
The mistake I almost made
After sending the first recommendation, I had 24 hours of doubt. What if the well pump started while the coolers were running? What if I had missed a panel schedule line? I called the owner and asked for photos of the breaker panel. That call exposed a subpanel feeding a 1.5 kW space heater. The continuous load went from about 4.6 kW to 6.1 kW, and the starting surge went up with it. A 7 kW generator would have been running at nearly full load for hours before the first motor even kicked in.
That detail, if missed, could have left the site without heat and refrigeration, and no relationship with the customer. The order was small enough that many installers would not have made the extra call. Small does not mean unimportant; it usually means there is less room for an expensive redo.
People often choose an SMA-inverter system for reliability. They will also quote SMA 2023 inverter output sold GW as proof. I use SMA products in many designs and I believe that proof. But an installed base does not remove the need for a site audit. A reliable inverter will not know that a client owns an unlisted space heater or a 7.4V battery charger unless I find it before sizing.
Short version: the decision rule I use now
The cleanest answer, after enough site data, is usually not one machine. It looks more like this:
- If the site only needs backup for short utility outages, buy a gas generator, store stabilizer, and include a transfer switch.
- If the site consumes power every day, or is exposed to long outages, put PV and battery at the front. Use a reliable inverter in the correct size, like an SMA 7000 inverter or another unit with documented overload behavior. The generator, if any, becomes the backup for the backup.
- If the customer is small, do not shrink the process. Do the same load study, ask the same questions, and calculate the same annual run hours.
I still install generators on some sites, and I still recommend solar on others. The lesson from the near miss is not that gas or solar always wins. The lesson is that the solar vs gas generator question cannot be answered until the loads, the start sequence, and the expected outage time are known. Small orders deserve the same scrutiny as large ones. The people who treated my first small project seriously are the same people I recommend now.
