If things go right 8 times out of 10 overall, that's a net win. Expected value - I get Product X $2 cheaper if things go right, but lose $6 if things go wrong. Expected value = 2*.8-6*.2 = .4. Overall net positive.
Once the initial investment in creating a stock of those items is done, there's only a minimal storage cost. For things like chips for cars, that storage cost is probably measured in double-digit dollars a year to create billions of dollars worth of vehicles.
Is investing in a business to smooth over supply chain problems really so taboo to a modern businessman?