You're getting into a BIG trap here:
Commitment Bias (Escalation of commitment) - The Decision Lab
Running a house plus an ADU off of a 100A panel is insane. Expecting Powerwalls to "save" you from overdrawing your main breaker makes the situation highly unpredictable. You do not have any control over when your tenants decide to cook or use the AC. Shopping for lower amperage appliances increases your "commitment" to this strategy further. Constructing an ADU is expensive; a service upgrade is peanuts in comparison. Furthermore, you're going to have plumbers, electricians, and all of the other contractors onsite already just to get the ADU built and plumbed and wired, and to install the solar and all of these Powerwalls. You'd get at least some discount on the upgrade because all of those people had to be on site anyway. If, after all of this effort, you're still tripping breakers, you're not only going to have frustrated tenants but you'll have to do the service upgrade anyway, at higher expense, and you'll also have replaced perfectly good appliances with useful life left in them (another expense). But at that point, you'll likely be in so deep that you'll try other strategies like pre-cooling your house or replacing your HVAC system, etc.
For the record, I believe a Powerwall is warrantied for 37800 kWh of energy cycled through it, and costs around $11k. That's 29¢/kWh in expenses just for the storage! I get that PG&E is a ripoff but the cost delta between off-peak and peak rates (ETOU-D) is only 4¢ most of the year and 12¢ during "summer" (June-September) and peak rates are only in effect for 3 hours/day 5 days a week..15 hours out of 168. Assuming the Powerwall assists you in avoiding peak costs, you'd have to consume 100,000 kWh during that time to have it pay for itself (except you can't, because it's likely that it won't last that long). If we assume that you do this over 10 years, there are roughly 7800 peak hours over 10 years on ETOU-D. Are you really going to be consuming 12.8 kW whenever peak hours roll around because that's what you'd have to do in order to hit that target. But wait, it gets better...the Powerwall can only supply 5kW max, so you'll need to extend that time frame to 25 years. Oops wait...that assumes you're drawing 5kW for 3 hours a day on weekdays. Can't do that either, as the Powerwall doesn't have 15kWh of usable energy in it. Gotta extend that timeline to closer to 30 years. But wait...we know that if batteries are cycled close to 100% every day, they won't last very long, so there's no way I'd expect the thing to even last 10 years under those usage conditions.
In addition, I expect storage costs to decline dramatically in the future in terms of what you're paying per kWh of storage while service upgrade costs will probably increase. In other words, I think you'd be best holding off on the solar and storage and spending the money on a service upgrade instead.
Look, I get that solar and storage is cool. It's the future. I get that people want to do the right thing for the environment. There's some value in that to a lot of people. But if you're a finances/numbers type of person, I just don't see how this stuff even remotely makes sense right now. The Powerwall is way, way too expensive for the amount of storage you get. The economics on EVs make a lot of sense, but the economics on Powerwalls, at their current prices, do not. As a comparison, the Tesla Model 3 long range has a usable battery capacity of ~75 kWh and costs about $55k. And you don't just get batteries for that amount of money, you get an entire car! In the world of Powerwalls, you'd be paying over $60k just to get 75 kWh worth of usable storage and that doesn't include the solar.