The power supplies you describe sound like an interesting option, especially if their MTBF proves to be up to the challenge. SMPSs tend to be small and light enough that installing two with a failsafe relay is another option. Please keep in mind that not all relay circuits are "klugy", and relays properly chosen for the application can have very impressive reliability figures, as borne out by their continued use in many industrial settings. Also note that the two-component relay-based warmup circuit I described is actually quite elegant: if the relay fails in any way (open coil or faulty contacts), the circuit will continue to pass current to the filaments. The components get ungainly in a big amp, but in a 12V string in a smaller amp that circuit is a good solution.

After reading your comment, I again checked for data and research regarding the relationship between tube filament voltage and longevity, but I couldn't find anything authoritative that supported your viewpoint. One paper found that running 6.3V filaments at 6.1V extended their life by 50%. Transmitter maintenance personnel often run new tubes at slightly higher than their rated filament voltage, then back it off to a slightly lower than nominal voltage to maximize their life. Keep in mind that 6.0V is less than 5% lower than 6.3V, and that filaments fed by a power transformer winding typically see much wider voltage deviations due to normal variations in the mains voltage (6 to 10%), historical changes in the mains voltage (5 to 10%), and normal variations in transformer windings. Also keep in mind that filament failure is a relatively minor cause of overall tube failures. Of course, there's no harm at all a running audio amplifier tube filaments at precisely 6.3 or 12.6V (and fully current regulated) if you can find a way to do that, but in practical applications this pursuit would soon give way to diminishing returns.