Apple making their own CPUs would also not have taken down the company. They took small incremental steps; at any point, a failure of their own CPU program would at worst have meant using a commercially available SOC (with probably great terms), or alternatively skipping the SOC refresh for a year.
In fact, by your criteria, I doubt that Apple has made a single risky bet this millennium. Even the iPhone failing would not have taken down the company at that stage.
Look, if you want some new hardware using latest node it’s expensive. You must sign nice expensive contract with foundry and pay pay pay before anything comes out. It’s risky, there is no guarantee, that all produced chips work as expected. And it’s done in secret for years with insane stress for engineers. They can leak or chip can be a failure. There is no way to get any SoC on the marked in desired quantities for Apple with short lead times. Changing processor was risky for them.
Which specific Apple-made SOC do you claim was a company-ending risk? The A4? Surely not, it was just commodity components they integrated. The A6? Clearly not. That's the first SOC where they included their own CPU design, but if it had failed they could just have used the A5 as a backup, given they were already manufacturing it on the same process as the A6 was released on. And so on.
Why do you think Apple didn't recognize whatever risks you think were there, and mitigate for them? Are they really so badly managed that they'd just risk a trillion dollar company on a roll of the dice rather than have the appropriate contingencies?
Why doesn't every single chip maker run exactly those same risks? Heck, why doesn't every user of commodity CPUs run those risks? The things you claim that could happen to Apple's SOCs could also happen to Qualcomm. Why isn't every user of Qualcomm's chips up the shit creek if a Qualcomm CPU fails validation?
I don't know where you got that from. You defined the criteria. By that criteria, neither Apple or Amazon ran a risky CPU project.
Now, the reason is that your criteria for what is risky are pretty stupid, as evidenced by not even the iPhone meeting that bar. But don't blame me for it: again, you are the one who set the bar at the failure of the entire company.
In fact, by your criteria, I doubt that Apple has made a single risky bet this millennium. Even the iPhone failing would not have taken down the company at that stage.