A guess: take Instacart.
Plan A: They know that people spend $x billion on groceries per annum in sf. They also know that a big chunk of that has to be very expensive real estate and relatively expensive employees who have to be able to afford to live relatively close to sf.
If you can build a warehouse an hour out of sf on dirt cheap land and recover the cost of grocery stores in sf, plus shave some off employee wages, maybe you can make a profitable business out of grocery delivery. This business will need a lot of volume to cover fixed costs. Why will they succeed when Webvan failed? More comfort with online shopping and delivery, cheap labor via exploiting underemployed people, including dumping infrastructure costs off onto employees by making them use their own cars.
Plan B: Google is in a death match with amazon. To first approximation, G's business is taxing ecommerce by owning discovery. If people start on amazon not google, google loses. So google is building out google express because if they don't, amazon ends their business. Therefore maybe instacart is strategic to google.
I don't think I believe A or B, really, but VC is a gamble. If someone can make it work, it's a $1.4T/year market [1]
Oh, and people don't earn billions by taking safe bets.
[1] http://www.ers.usda.gov/data-products/ag-and-food-statistics...