I am aware of the potential issues with a new game, but it's the only solution to solve most if not all of the pains of Odyssey / Horizons. Also don't forget the economics of this.
What they need to do, in order not to alienate the current player base (myself included) is to allow your current CMDR to be "imported" in the new game. Imagine a UE5 generation engine (DX12/Vulkan, RT, etc.) Elite game, with fantastic graphics and decent performance for the hardware, with all main issues solved, re-written from zero, but benefiting from everything that's good in the current game and more... and with your own current CMDR in it. Wouldn't you pay 60 USD/EUR for it?
But will Frontier be able to do it? I'm skeptical of that.
Sure, imagine a game like ED but better and I'd buy it even if I couldn't copy my ED CMDR.
The thing is, both the economics and the alienation issue are
much easier for anyone except Frontier to solve.
- if it's being developed by another company entirely, it's not competing with support for Elite Dangerous. If ED gets put on hold entirely for five-to-ten years while ED2 is developed, no-one's going to care what it's like by the end or whether they can import that dusty ED1 account from years ago. If ED doesn't get put on hold they're now trying to fund development of two entirely different but competing games and can't "release" ED2 until it's got all or at least most of ED's ever-increasing feature set.
- if it's an entirely new game by a new company, no-one's even going to seriously consider importing ED or Freespace 2 or Space Invaders savegames. Any game which fixes most of ED's serious issues while still being close enough to ED-style gameplay ... isn't going to have a meaningful commander import option anyway.
- if it's not being developed by Frontier, the fact that Frontier isn't capable of developing a better ED-like game than Elite Dangerous doesn't apply either. Admittedly most other companies aren't capable of doing that either, but maybe there's some.
- better graphics is if anything exactly the thing ED doesn't need, because modelling and texturing to the quality level required even by the current engine is already a really obvious bottleneck for them. Deliberately moving away from photorealistic style to something more stylised and easier to draw (perhaps easy enough to partially hand-off to procedural generation, even) would probably be a better move.
None of ED's big issues - slow and often-delayed release cycle, underwhelming headline features, core gameplay based on 40-year-old hardware-limited designs - are going to be fixed by Frontier having another go. But they've also all been really obvious since 2016, so if anyone else was going to bring out an actual competitor ("you can also fly spaceships in this game" isn't sufficient) they could have done that already.