Sorry, but the math doesn't support your view that Intel was better off spending $15B on this acquisition at this late stage versus actually having an innovation program internally or purchasing a whole slew of early bets (most of which don't pay off.)
MobileEye today only has about 500 employees, of which maybe half are Engineers. It's likely that only a small team of ~30-50 was involved in building out the majority of the technology that makes up the core of its IP. With $15B, Intel could have funded 200 fifty person teams, each for a full decade. Given that the investment would have been done over time and various investing and culling strategies would have been employed, even with a very small success rate, I contend that Intel could have researched their way to early leadership in the autonomous vehicle market and maybe a dozen others with the same level of investment they poured into MobilEye.