The beginning is nearly always federal R&D funding. Much of it won't work, sure, and that's fine. It's not wasted, because when it works, it creates such a massive everlasting surplus and opportunity machine that it overcomes all past failures by orders of magnitude. Such as, computers, and all they enabled over the last 100-ish years.
The myth of the lone inventor in the garage should have been updated even in the pre-WW2 era.
Alas when the govt follows the exact same model, taking high risk, high reward bets, then it's seen as "wasteful spending". Despite the staggering value of the wins, it becomes better to "spend nothing" than waste a penny on research that goes nowhere.
The levels of cognitive dissonance, not to mention hypocrisy, are truly incredible.
And the charge is being lead by someone who literally made his wealth from this model.
ARPA-E is high risk.
ARPA-H is high risk.
Much of NSF is high risk, like NSF Engines, NSF Future Manufacturing, NSF Convergence Accelerators.
DoD SBIR/STTR is high risk. (Confirm it for yourself and look at this month's topics in https://www.dodsbirsttr.mil/topics-app/.)
AFWERX is high risk.
SpaceWERX is high risk.
DIU is high risk.
NASA SBIR is high risk.
NASA NIAC is ultra-high risk.
DoD Office of Strategic Capital is high risk, the kind of risk no investors would fund.
Investors scrutinize pitch decks and then do hard company due diligence which frequently falls through. And conversations die-off with no obligation to provide feedback, unlike in government. And investors will not fund true R&D. They fund scale.
So no, your statement does not hold.