They're in a better position to take big bets than anyone else - Google X, Microsoft Research, Facebook's FAIR and others. If they have the right people in house (or can get them), they can take big swings.
Public companies often can't take big risks because they have to justify them to their shareholders and consistently meet expectations. The examples you've listed are all just R&D projects, the expense of which are a tiny line item in the budgets of the companies.
Conversely, a startup has nothing to lose. Early-stage investors know they're taking a moonshot, and they expect only a few of their bets to succeed, so they will encourage the team to swing hard. That's how startups can disrupt established markets with big players: they have less to lose, and don't need to earn buy-in of a large existing company. They can build what they need as they go.
This is also why most acquisitions fail. The team of the acquired company can have difficulty navigating the existing company culture. Often, they're tasked with being a "change agent" only to be identified as a threat by the acquiring company's "immune system" (its existing power structure).
Well, they can take risks, but new projects have to make sense to their bottom line. If you are 10B company and you are looking on 5 - 10M opportunity than it doesn't take much sense for this company. There is great book on this topic called Innovator's Dilemma.
They're in a better position to take big bets than anyone else - Google X, Microsoft Research, Facebook's FAIR and others. If they have the right people in house (or can get them), they can take big swings.