SET 45 and their (real) required speaker efficiency

VVCV, There are several factors that determine the speaker sensitivity needed to use 45 tube amps. Many have already been stated. Room size, music density, max SPLs for dynamic peaks, average listening volume, etc.Keep in mind that many manufacturer's speaker sensitivity ratings are for one central frequency. As ddriveman suggested the lack of headroom for dynamic peaks is often the limiting factor. Speakers without crossovers can help. Passive crossover can sometimes be a more difficult load and can consume power. A flat impedance curve and minimal phase anomalies help.

I use mostly low power SET amps to demo our speakers which have a sensitivity of 105 dB. We have two pair of 45 mono amps, a pair of 2A3 mono amps, a pair of 300B mono amps and a few other amps of various types. Our showroom is 21 feet by 31 feet with a ceiling that starts at 9 feet high and goes to about 14 high feet at the rear. Our speakers have separate powered woofers so the SET amps don't have low bass duty. The main towers go down to about 100 Hz but we typically set the active crossovers to 120 Hz. The amps are directly coupled (no passive crossovers) to the drivers. The 45 tube amps, to me, sound the best and have no trouble driving the towers to high volumes with no compression. In our setup the 2A3 amps are a close second. Our speakers are a line source so the dB loss per distance is less than a point source speaker which helps. 

Our speakers are unique in many ways so maybe not the best example but typically for a medium size room, listening to reasonable/enjoyable volumes a minimum sensitivity of 100 dB should work. With so many variables you should audition the potential combos.

@vvcv  A nice plan B to consider:

If you are set on the type 45 power tube, why not get an amp that uses a pair in push-pull? I've done direct comparisons; the thing about any SET is that you really don't want to run them past about 20% of their total output power, since higher ordered harmonics start showing up on the transients (where the power is). This causes the amp to sound 'dynamic' as the ear uses the higher ordered harmonics to sense sound pressure- its distortion masquerading as 'dynamics'.

But with the same tube operating in push-pull, but otherwise class A and zero feedback, the total power will be more like 6-7 watts and about 90% of that power will be usable as opposed to 20%. So you will have a much wider range of speakers you can play.

Now to the comparison; side by side it was obvious that the push-pull amp using 45s was more detailed (vocals were easier to make out) and had more bandwidth both in the bass and the highs and every bit as smooth. IOW, no downside except you need a pair of power tubes per channel.

From a great deal of experience listening and comparing different amp types on 2 way 12 ohm 96 db speakers I couldnt even live with the output of a parallel single ended 300B amp. Ralph (Atmasphere) is 100% correct in that these amps fool you into thinking that the output is adequate when it is actually colored as hell. What you are hearing is not dynamic nor do I think it is correct. This at least has been my experience over several hours of comparisons per day over several days with numerous amp designs. On high impedance speakers the clear winning design has always been an OTL. 

Had a pair of LaScalas powered by Don Allen's 300B/ 2A3 SE switchable amp. Also a Don Allen SE EL84 amp.They could rattle the windows in the LR. And play nice and sweet and quiet as you would like.

Put a higher wattage OTL on your LaScalas and see if you feel the same way. Of course the presentation will be much more linear and much less colored which I appreciate and others dont.