One thing to remember is a big reason new tape machine manufacture became unprofitable is because old tape machines keep on ticking. It’s much more cost effective to buy used and restore to like new condition than to pay new prices. Manufacturers were competing with their own legacy equipment.
I think too much of the music "talent" of our time have come to expect and rely on the many vices of digital audio production, rather than the more austere, demanding process of tape. So the market is still relatively small and probably best met with restoration of old machines. To seriously consider a new machine, there would have to be more of a market for use of the machines.
Even then, it will take a genius who can follow in the footsteps of John Stephens to come close to making a buck selling machines to pros. John made arguably the best pro machines ever, and certainly the most portable by far, and he made a number of changes to the conventional designs of his era that must be followed on and extended to have any hope of a commercial product.
It must avoid any physical component that can be avoided, to cut cost, size and weight.
John designed his own, small, proprietary electronics to slash cost, weight and space. He used unbalanced audio without transformers to save an immense amount of cost and weight and get cleaner sound in the deal. And he designed a capstanless, servo-controlled transport, eliminating capstan, pinch roller, capstan motor, etc.
With modern control circuitry, a servo-controlled, capstanless transport should be easy. With digital controls and surface-mount electronics assemblies, the record and playback electronics should be self-regulating, small and robust. Probably use DB25s for I/O on a multi-track machine, quickly connecting with snakes to the studio patchbay or to a portable rack for live/remote use.
Obviously, there could be use for new, high-quality 2-track machines. Harder to assess how many of the multitrack formats would make sense. I tend to think fewer, maybe 8 and 16, recognizing that anyone interested in more tracks with probably want ridiculous track counts on a DAW. OTOH, I could be wrong. Stephens sold a number of 40-track 2" machines, some of which are still in service. AFAIK, though, any number of tracks on a complete, new 2" head stack is going to be an expensive chunk.
Just my ramblings, and more than I can or wish to undertake, but if the circumstances were right, the right consortium of talents, time and money could do something exciting.
Cheers,
Otto