Then the engine really does have to be a problem. Because that will be 2+ years worth of work with quite a few resources on it to produce like ten hours of main story content and four-five hours of side content. That would be pretty pathetic in terms of time spent vs. content produced.
I feel a need to comment after much on this subject I've read here. It's not specifically aimed at you in it's entirety, nor is it meant as an attack; your comment was simply the last straw.
An engine is only a problem in that resources required, ie people time and money, to support it outweigh it's effective use. The reason some things "just can't be changed" in a system is that the team(s) lack the intellect and/or time to do so. It's why some issues often persists in a studios games from release to release. It could be lacking visuals, controller/keybinding options, or even the ability to exit the game without having to load the title menu once again. Often it's caused by time compounded on lack of skill. When you're hustling from project to project, being overworked, *in the case of Asia specifically* being underpaid, you lack the time or even willingness to really dig in beyond what's required unless you have a passion for that certain element. Some issues are considered so small that the work to fix them may not be worth the costs. I don't mean just upfront man power/money, but the impact to other projects. I'm going to prattle on for a minute about why an engine is almost never bad per se.
An example in CDPR's case - AI. Mediocre at best. They just are not good at AI programming, or at least getting it to work efficiently. These systems are well known. Sometimes they're not necessary for a game, it's arguable about what's necessary for 2077. The issue is they really talked up this side of the game. Until months before release, even as they were still making changes, finishing features, making the game turn on, they bragged about their AI and police system(s). Why does this game have bad AI after a year of nearly absent AI though? It's likely that it wasn't working properly and they lacked the skill to make it work on top of all the other broken systems. There was a 'better' police system at one point, there were at least plans at better ped AI, and I doubt they want city infrastructure to come to a complete standstill when you sit in the middle of the road. This still happened every time the last I played the game. Switching engines will not fix it. I would still lean toward the notion I've carried that they prioritized fidelity to such an extent that other systems suffered. Even at release, with all the problems the game suffered across platforms, that it could run at all is notable. The game, still being a bit of a mess and lacking marketed features, being able to run and in many cases run well is an achievement worthy of praise. REDengine is not a piece of crap. If they can get good, even decent, ped/drive/police/enemy AI all working at once on top of everything else - it will be absolutely lamentable that they're throwing away such impressive technology. It can be replicated though if they manage their resources properly in the future.
An example of this in terms of fidelity - Fromsoft games, most of which I adore, appear as a previous gen title at release in terms of visuals. No exception. This isn't simply 'da engine is bad', the programmers simply lack the time or skill to make wowy graffix work efficiently. Almost any engine can do the fancy rendering, artists can place the lights with grand effect, and the tech art side can write juicy shaders easily. They've shown the ability to make fantastic rendering with the Lord of Hollows ending in DS3, but they aren't capable of doing that at real time it seems. Shadows and rendering seems to be what really keeps their games from having that grounded look. The only title of theirs, which still looked a bit dated at release, where these things weren't a persistent issue is Sekiro. The community has often propped up the idea that if they just change engines their graphics problems will be solved, and we'll finally have a visually spectacular game to truly create the atmosphere proposed by the content. Not true. Fromsoft made a UE title a few years ago. Guess what? It did not look visually impressive. It didn't look 'bad', but nothing special in pure graphical terms. It looked old, again, just like all of their other games. Gameplay and Aesthetics over fidelity any day for me, but objectively they lack in the graphics department. They've stated this openly themselves.
These long winded examples are not insults to the studios or their teams, they're intended to make the point that the engine is only as good as the people working it.
Those impressive videos of one-man built UE5 "games" on youtube? They're asset flips where no one can do anything. Start adding proper physics, physics cancellation areas, interactive objects, character customization, vehicles, mounts, decal systems, ai, ui, sub ui, consumables, inventory, etc, etc. These captures of "gameplay" don't mean much of anything really. When multi-billion dollar corporations put out fake demonstrations, how real do you think some random dude's game on youtube is?
The best reason to move to UE5 is management has driven the legacy developers away - people with deep knowledge that could really get things going aren't there. Now you have a lot of people with UE experience in the market, the engine is complex yarn of systems on top of systems(before and after 1.5 I made long post about how the entire rendering pipeline from release until February 15th was out of order. It was objectively utter trash), so it's a no brainer to switch. That's sad, but it's really the best case scenario because it means there is potential to recover based on management's future decisions. The worst reason would be that people at the top of the food chain saw the Metahuman and Matrix demo, paused, then thought "Wow, let's just use that." not considering that a demo of an uncanny valley person is room with perfectly staged lights.
TL;DR - An engine, and games created with it, is only as bad as the studio allows it to be.