That’s fair, but come to think of it, the architecture of the CPU doesn’t really say anything about privacy. Someone can build an RISC-V chip but sneak in telemetry, or you can build yourself a x86_64 CPU and be 100% no telemetry. It’s about the manufacturer, not the architecture.
I don’t think you can ever be 100% sure that the CPU you’re running on is telemetry-free unless you have those kick-ass X-ray machines and examine it yourself. Building your trust on top of something else you deem trustworthy though, is practical. Billions of people are running Intel/AMD off-the-shelf CPUs, and there are perhaps millions specialists in them, what is the chance that a backdoor remains hidden?
The same goes for software. How do you know Linux kernel, OpenSSL, Wayland is trustworthy? Because many people use them, and it’s unlikely a backdoor is there. Think about the sheer amount of software the CPU runs. Don’t you think we shall have a greater concern there?
Hopefully this calms your paranoia about hardware a bit.
I’m that goof who built a desktop-shaped computer with the Milk-V Jupiter RISC-V (RV64GCVB) board last year. The performance was awful compared to recent amd64 setup. The amount of RAM wasn’t an issue, but the CPU was way to slow to even run a graphical interface (read XFCE), or compile some simple program in Rust. It’s equivalent to those arm926ej-s SBCs ten to fifteen years ago capable of running Linux 2.6, and only suitable for writing code in text.
I’d say give it two more decades, at current pace.