So even if Switch 2 doesn't make it all the way to 2035 with zero cracks, there's a strong likelihood that any exploits found will be short-lived.
Anyway, situations like the one you describe are one to be solved by legislation requiring certain devices be sold as open devices that put power in the hands of the owner.
My PS3 and PS4 were both jailbroken/rooted. I don't remember the ps3 routine, but the PS4 was loading the "system -> help" page while connected to a ESP32 wifi AP running a simple web server that replied to requests with the jailbreak for PS4.
I give it about a year, especially if nintendo has to change the specs or otherwise tampers with customer expectations. there's bound to be some way to reload firmware on a "dead" device without pulling chips, and that's all it takes.
The two pins were installed by design from Nintendo to activate the Tegra RCM mode. RCM mode meanwhile has a USB buffer overflow which is the real bug.
In modern NVIDIA chips, this RCM mode no longer exists. The new recovery modes meanwhile are running across multiple physically separate CPUs verifying each other (glitch one, the other notices), all running formally verified firmware written in SPARK (the thing you use for nuclear reactors and avionics).
As for the OS itself, according to a maintainer who rewrote the kernel twice for open source, it has zero bugs. None. The microkernel is tiny, has no drivers, and almost no attack surface. This is born out by WebKit exploits being a dime a dozen on Switch, but all of them are useless.
I guess that, when you absolutely want zero surprises, Ada is the language of choice.
This is hyperbole. We have 1 switch that routinely "won't power on" without a ritual of button holding & timing. My original switch used to hard lock, but i stopped trying to play the sorts of games that were causing the OS to crash.
Both of these disprove the zero bugs claim, unless we move the goalposts.
That's like saying "I plugged in my phone's charging cable, and unplugged it, 20,000 times, and now it's sometimes showing the charging symbol inconsistently, obviously a software bug proving the charging circuit driver has a security flaw."
Even if it were possible to find a vulnerability in the hardware, doing so without attracting the attention of law enforcement will be profoundly difficult, as Windows sends telemetry back to Microsoft about every instruction that runs on your hardware. Apple will claim to be more privacy-focused, at least for a year or two, but the M9 chip's NPU will just perform local inference on your activity and report you to Apple and the FBI if it detects attempts to break security.
Maybe we should think about this like the concept of public domain. Locked down for X years in order to protect the artist, then opened up for everyone to benefit society.
The proof in the pudding of this will be when the Nintendo Switch 2 reaches 2035 with no cracks. That's my prophecy; that this time around the cat actually will catch the mouse. Between NVIDIA's heavily revised glitch-resistant RISC-V security architecture and Nintendo's impeccable microkernel, there's nowhere left to hide. DRM may turn out to have been a very slow long battle to "victory," not a "this will always be defeated."