The calculations say 20 batteries would run great for at least 9 hours after dark.
Theory and practice are the same in theory, but they are different in practice.
The deeper questions are:
What is your budget?
How much risk are you comfortable with?
What is your technical expertise?
Who can you rely on for the expertise you don't have?
The research and budgeting phases should precede the buying phase.
One machine failure out of 10 machine is unfortunate. But having only one machine and see it fail really sucks.
10k solar will power at most one unit (any S19 model) for reasons already mentioned by others. Then it's fun and games until the thing fails. Then it may or may not be possible to fix. Or it may last long after breaking even.
I think such a failure is relatively unlikely, but I would sleep so much better going from 5 functional units to 4 than from 1 to zero. In that sense, multiplying the setup may be "less risky" in that way.
Now, taking the "10kW one unit" risk knowingly, in a calculated and conscious way, is totally fine. It's better to know what could happen from the start than learning it the hard way.
If I barely had the budget for a 10kW capacity and a single S19, I would not take the risk. Even with no technical issue, it may still never break even.
If your goal is more about getting your hands dirty, learning and having fun than about return on investment, that's counts for a lot.
Have you used some profitability calculators online?
Did you precisely calculate your theoretical cost per kWh over different time frames?
Can you:
Buy and building the setup,
Look at the nice flow of income for a few weeks with a big smile,
Calculate that you expect to break even in 6 month (and so rich in a few years),
Smell some smoke and melted electronic (not covered by warranty or insurance), AND
Still be happy to have tried?
Also, miners that are breaking even at night and very profitable during the day is not too bad.
If the choice is between mining and buying BTC straight up, again the approach is different.
Keep in mind 10k will run AT MOST one miner IF and only if the theory and the environmental conditions all work in your favor.
Also, batteries have a life cycle, and solar panels don't last forever either.