It's interesting you mention that, because in reviews for the STAX SR-009 or Abyss AB-1266, I occasionally see phrases like "You'd have to spend hundreds of thousands of dollars on a speaker system to get better sound than this." (Not my claim.) For me, it's a moot point, because I simply prefer headphones over speakers, so from my perspective, it's not so bad. But I definitely see your point how many people would want to go for speakers over ultra-high-end headphones.
I enjoy both for what they are good: headphones for the intimacy and pureness of the sound, and speakers for realistic sound stage in a real physical space (short time ago I would have added bass impact here, but not any more). Still speakers have the advantage to produce sound waves crossing your whole body - if that adds to the excitement
.
About pricing: in high end audio, you pay for the price. It's about perceived value.
As with speakers, most of the real cost is coming from R&D, and the limited number of units sold. It's very hard for a small company to estimate pricing so that they are on the safe side. However, prices tend to creep up too much nowadays because the perceived value is bubbling up.
When you compare against the HD650, you price accordingly the HD800. When you compare against the Orpheus, you price the Stax 009. Then you price the Abyss against the 009, then the HE1000 against both, and the LCD-4 too. Remember they have been relative all along. If Sennheiser brings out the Orpheus successor with a 5 digit price, all prices will rise, since you compare the diminishing returns against something that costs 5 figures.
The sound quality of today will be available for much cheaper tomorrow, and the results are sometimes step- or jump-improvements, but then mostly incremental for a long time. The HD800 made a big difference on technology, sound quality and price level. Today it's a bargain. The Stax 009 is the absolute reference for neutral musicality (for me), i.e. a HD800 done right, but it should not cost much more than the HD800 when judging the build quality, the used materials, and the R&D costs. The HE1000 strikes a good balance between natural sound and musicality, but no way justifies its asking price, even though there is more material and innovation than with the Stax 009. The LCD-3 was pretty much as expensive as money can stretch for a headphone, overpriced on its own, and I bet it's possible to tune it towards a more open sound balance and then how far would it be from the king-to-come LCD-4? As for me, I am happier with my modded Stax 007 Mk1 than with the Stax 009, HE1000 or LCD-3 - so the diminishing returns will more and more frequently hit people's threshold for "good enough to enjoy the music" quality, even with older technology.
Much of the improvements/diminishing returns are sunk in the mud created by other audio components: microphones in the big part, then the mastering/authoring chain, sound bearer (noisy channel), pickup/decoding device, amplifier, etc., then your golden ears and your grey brain.
IMO a truly ultra high end system should address all the elements of the audio chain, and use synergistic solutions for transferring acoustic information. For instance, I would apply some of the learning done in microphone tech (some really expensive) in headphones as well, use similar materials across the audio elements, etc. These would also bring total cost benefits, making it possible that a relatively mid-price chain would put in shame many high end systems of today with less synergy (it's another question how much of these savings end up visible for the buyer).
Since most people have a relatively established threshold for "good enough to enjoy the music", it's a good strategy to find the cheapest chain of old (second hand) tech that causes the much-desired goosebumps, and let the newcomers be roasted a little bit for perceived, and relative value.