It's relatively common to find residential inverters that handle the balance issue because they interoperate with utility power management and will constrain themselves if told to do so. Inverters that handle battery and solar input are used here to create an ad hoc peak power plant. National Grid would pay me $500 to $2000 a year for providing extra energy into the grid during peak loads.
You also see this with utility-scale solar, where the power consumed from the solar flat tops being the access is dumped. I take the position that if your solar array is constrained during the day, you need more batteries.
Vermont has a virtual power plant where solar panels are installed on residential housing, batteries are supplied, and the virtual power plant is managed. As far as I know, they have eliminated the need for a gas peaker. Tesla does the same thing as Sunrun.
As I often say, your objections are engineering problems that can be solved. :-) If you want to go further down this rabbit hole, check out https://www.youtube.com/@WillProwse. There is lots of good information, good testing, and a willingness to trigger circuit breakers at the 100amp+ range or let out magic smoke.
In some of his best videos, he cuts open battery packs for various manufacturers and evaluates manufacturing, sensor placement, BEM, and hot and cold temperature protection issues.
One thing we did not discuss is the need to have everyone connected to the grid contribute to the maintenance of the grid. I've lived in communities with municipal electric companies, and they generally have more reliable infrastructure and cheaper power than the big boys like National Grid and EverSource. Done right, we could use Agravoltaic to help secure the financial footing of local farms and improve many aspects of farming, including soil moisture retention, better yields, etc. With goats and sheep livestock, you don't need to mow, and you get a nicer local lamb source than you would get it by shipping it here from New Zealand.