Bleeding-edge == unpredictable?

Most IT projects don’t make it on time, and neither does the Tesla:

Well, five years after its founding, Tesla has shipped about 70 electric roadsters, and the car does in fact turn out to be a classic Silicon Valley product—it’s late and over budget, has gone through loads of redesigns, still has bugs and, at $109,000, costs more than originally planned. Tesla’s first 40 roadsters went out of the factory with a drivetrain that needs to be replaced. (Tesla will do the rip-and-replace for free.) Its second car, a sedan, has been delayed until 2011.

Perhaps the point of bleeding-edge technology is that it is not well understood and hence impossible to plan for?