There’s absolutely an extra switch that needs to be installed, but I think it’s not so much the transfer switch requirement but two main things; 1) the battery

There’s absolutely an extra switch that needs to be installed, but I think it’s not so much the transfer switch requirement but two main things; 1) the battery cost, and 2) net metering. Grid tie solar with net metering gives you all the benefits of a battery (except blackout coverage) with none of the cost. In other words, you get full retail value for 100% of your solar generation no matter when it comes or how much power you happen to be using at the time (because the meter runs backward, the grid acts like an infinite perfect battery) Net metering is a nice solar subsidy while batteries where extremely expensive, but as battery costs plummet I assume net metering will also disappear. It’s not really fair for the utility after all to be paying solar customers retail rates for their generation. Without net metering you will only get paid for solar power that you happen to use while it’s being generated, or that you can store for yourself to use later. Finally, I’m not 100% sure but I would imagine a solar system with a transfer switch and no battery will just immediately overload or brownout if the grid is not there to keep the voltage steady. I would assume you need a battery to be able to serve any variance in your demand (e.g. a compressor turning on) even if your 1 minute average demand is actually running below your generation capacity. TLDR: Batteries need to be cheap and net metering needs to die before you see most solar deployments that can run off grid. The good news — if batteries are durable then a battery backed grid tied system can actually pay for the batteries and then some, if you can arbitrage the demand pricing curve. This is a win-win for everyone (including utilities) and IMO absolutely the future we’re heading towards.