The distinction between "possible" and "economically viable" tends to collapse in practice. For hardware like this, people usually mean something closer to "no known attack within a realistic budget and skill envelope."

Voltage glitching has been around for a long time, but applying it against a tightly constrained boot chain with limited observability is a different class of problem. You are essentially searching a high-dimensional timing space with very little feedback. That is where most prior attempts seem to have failed.

What changed here is less the existence of the technique and more the instrumentation and persistence. Once you can reliably characterize the system’s behavior at that level, "unhackable" turns into "not yet mapped."

I think the safe analogy still holds if you interpret it as "no one had a stethoscope sensitive enough until now." The underlying weakness was probably always there, but practically inaccessible.