Getting good returns from a 100 billion dollar investment is hard. Just as an idea of scale that's enough money to more than double world wide fusion research spending for the next 30 years. Suppose they started down that path and 20 years from now started building useful and highly profitable fusion reactors that beat coal power plants. Now what if that failed. Measuring the risk / benefit curve on such an investment is hard. But, Apple could afford to fund 5 other projects on that scale without touching their cash horde.

IMO, Apple starting down the dividend path is simply the only reasonable course when faced with that sort of cash flow.