This isn't a technical limitation though, the European standard for airplanes newer than 2021 is in fact 25 hours [1].

[1] https://mentourpilot.com/who-doesnt-want-25-hour-cockpit-voi...

I don't think the problem you're describing is actually a problem.

Exposure to super-high temps occurs in a small set of circumstances, all of which overlap with the destruction of the recording device and the cessation of incoming data. So we only need the same 1.2GB (or whatever) of high-temperature-tolerant storage.

The 25 hour storage can be on normal flash, as if we're more than 2 hours past the incident and data is continuing to come in, then the incident of interest did not destroy the airplane, and the flash will have remained within its normal operating parameters.

Multiple investigations in the past have recovered data from FDR and/or CVR after an extensive high-temperature fire. I do not think that FAA will give that requirement up.

Yes. As I said. The existing system can remain in place, with all of its existing high-temperature-tolerant components.

In addition to not giving up that requirement, we could also add a longer, not-heat-tolerant storage. If it gets destroyed in a fire, see the above paragraph. If there is an incident where the data is of interest and the aircraft is not destroyed in a fire, then this will maintain the data long after the above system has deleted it.

No one has advocated giving up the high temperate storage.

I often wonder why we still rely so heavily on local storage when in-flight Internet exists. Flight data could be streamed in real time to the cloud for redundancy.

Read it more carefully.

What you described is not a data retention problem at all.

It's a material science problem, and other forms of media are affected by high temperatures and physical deformation just as much as flash if not more.