I saw this elsewhere and responded in another thread. The 17% appears to be *all* passive funds (many of which already own TSLA), not just the S&P funds. However, they give a number for S&P ETFs alone which amounts to 2.5% of the S&P market cap, and there are also S&P mutual funds (probably as much money in mutual funds as there are in ETFs, which would bring us to 5%). So maybe there are a lot of really small S&P funds which didn't show up when I tried to add them all up. And then there are still closet indexers.
So I guess the percentage is somewhere between 2.5% and 10%.
Since the S&P is market-cap-weighted, (money in S&P funds)/(market cap of S&P) = (amount of TSLA S&P funds have to buy)/(market cap of TSLA), which makes it simple: that percentage is the percent of outstanding TSLA shares the funds have to buy.
In answer to the other question which I didn't answer, the true really-tight S&P tracker funds have to buy it very quickly -- on a specific day no less. Other traders will, however, front-run the S&P addition by buying stock before the S&P addition in hopes of selling it to the S&P funds at a higher price (this is perfectly legal).
So most S&P funds give their managers leeway to do this "early purchase", or the "early sale" of the stock being removed from the S&P, themselves -- since most S&P investors are OK with their fund doing better than the S&P, and not so happy about doing worse than the S&P. And by definition a perfect tracking fund does worse than the S&P due to fees. Anyway, they'll probably buy it over the course of a week, maybe even more.
Because of the managers front-running S&P additions and the "closet indexer" funds who officially are actively managed but basically follow the S&P, the stock price effect of the S&P addition will probably happen over the course of months before the actual addition.