A single S9 probably would not make sense. Keep your mind free of the hassle. For a Bitcoin mining business, 10x S9's or equivalent is absolute worthwhile minimum, unless it's just a hobby. As general advice, you should go with Ethereum mining instead. As Ethereum mining has been more profitable than Bitcoin mining for 4 consecutive months in a row, it just makes more sense as an investment opportunity for virtually anyone who gets started with crypto mining today, or is at a point of having to replace legacy hardware. I'm getting my own feet wet in Ether mining just this week.

In regards to S7's, I know an investor in wind turbine farm business. He bulk buys S7's on second hand market for peanuts, and then feeds the excess electricity into S7s when production significantly exceeds demand. He says its "much better than selling it back to grid". Your observation is correct, for mining all year around, their worthiness is questionable to say the least. At the end of their life cycle. Perhaps scrap metal for good in 12-18 months.

And then there are other interesting and worthwhile hybrid mining business models where older units or low quantity of miners could make fiscal sense. I don't know much about such hybrid models, but there are now a number of booze makers who find valuable use for the heat generated by crypto miners. Ref: https://news.bitcoin.com/south-florida-distillers-bitcoin/