Well, my prediction since about a month after Civ5's release is that Civ5 will end up like HoMM4. There are many parallels:
1. Both games started their design process with the notion that the previous game in the series had reached that series' pinnacle of complexity and that a new approach needed to be taken in order to attract new customers.
2. Both games departed radically from their predecessors in ways that led to substantial design problems: In HoMM4, the "hero group"-armies, the hero-less armies, and the new hero evolution paths wreaked havoc to game balance; in Civ5 1upt required a total rebuild of the game mechanics, of which many ended up broken.
3. For both games, these changes led to an AI that was substantially weaker than that of the previous game, which led to a lot of criticism. For both games, the predecessor's AI was considered to be the best so far in the series, while the AI in the new game was seen as a severe step backwards.
4. In both games, in order to have enough time and resources to design and implement these radical new concepts, a substantial portion of the predecessor's content was left out of the new design. HoMM4 was the first HoMM game ever that had less factions and less creatures than its predecessor. Civ5 lacks religion, espionage, many trade and diplomacy options, meaningful improvements besides trade posts, etc. Both games were perceives as substantially less complex than their predecessors.
5. In both games, previously free systems were forced into arbitrary decision trees. In HoMM4, players suddenly needed to make exclusive choices about which creatures to train in their cities, choosing one path made the other path permanently unavailable for that city. In Civ5, the previously free and complex civics system was replaced by forced decisions about the non-revokable, permanent focus of one's civilization. In both games, this decision was heavily criticized and was not grounded in the game's atmosphere and feel, the change was seen as an unwelcome artificial limitation of the players' options.
6. In both games, lots of resources went into creating a new visual presentation of the game, without much success. HoMM4 marked the transition from (extremely well done) 2d graphics to new 3d graphics, and the new graphics were clumsily and amateurishly done. In Civ5, a new engine was programmed from scratch - as a result, the game's visuals struggled in many respects, and weren't seen as much of an improvement over Civ4's (less animations, amateurishly looking rivers on release ...).
7. Both games were rush-released in an atrociously broken and unfinished state. In both cases, the publisher of the series had changed, and the new game was the first of the series over which the new publisher had full control.
8. For both games, large amounts of the fanbase lost hope with the new game and went back to the previous one. In the HoMM fanbase, only a very small minority regards HoMM4 a better game than HoMM3. Civ5, in recent polls on CivFanatics, trails far behind Civ4 and sometimes even struggles securing the third place. The Civ5 forums consistently have less views than the Civ4 forums. Civ4 sales have resurged after the release of Civ5, and newcomers often get recommended to buy Civ4 instead of Civ5 unless they don't want their game to be very complex.
9. For both games, the modding community never reached the heights that it did in the previous games. HoMM4 mods never added much content, the largest mod for it is a balancing effort which addresses many issues that never got solved by the official patches - whereas for HoMM3, modders added tons of new content and game mechanics, including a new level of creatures, a (previously non-present) experience system for all creatures, new abilities, and even a whole scripting engine. For Civ5, we cannot say conclusively (yet) what modders will be able (and motivated) to do with it, but so far, the modding community is developing very similar to that of Homm4: Many people losing hope, or finding that the game just isn't worth the effort to mod it extensively, and the modders' main focus is fixing broken mechanics instead of adding new and more complex ones.
If this analysis is correct and if there are indeed many parallels between HoMM4 and Civ5, then what can we expect for the future of Civ5? Well, the prospect is rather bland. HoMM4 received enough makeshift fixes to make it playable for the people who liked the new approach, so that expansion packs could be released. The fundamental gameplay issues that made the game worse than its predecessor were never addressed. The publisher declared these new ideas to be the next best thing since sliced bread, as long as the game was being sold. Later, when the marketing for HoMM5 started, the same publisher made an effort to point out how different the new game's mechanics were from HoMM4 and that many "missing" elements from HoMM3 would be reintroduced. The effort was only partly successful. While most players enjoy HoMM5 far more than HoMM4, many still regard HoMM3 as the best of the series. On a small scale, modding for HoMM3 still continues. New factions are being developed, and a promising effort to write a new (open source) engine may finally overcome the limitations that the original engine has. Personally, I'm more excited over new releases of that engine than I am over HoMM6. I guess I'm not the only one. HoMM6 might end up being an enjoyable game (and I'll probably buy it in a year or two, when prices have fallen and the public beta / release is over), but given the direction the publisher has taken, it's unlikely to have the depth that I'm looking for.
My prediction for Civ5 is that it will develop on a very similar course. The fanbase will continue to favor Civ4, while the publisher will call it a great sucess, until the day when the marketing for its sequel starts. From this point onwards, we will hear how much Civ6 "improves on the shortcomings of Civ5", which are suddenly acknowledged as if the publisher hadn't hyped them for years. Civ6 will not reach the complexity of Civ4; instead, in an effort to re-attract some of the old fans while not losing the new ones, it will be in the middle between Civ4 and Civ5. Many hardcore fans of the series will still prefer Civ4, for which a few mods are still actively developed. Civ5 will mostly be seen as a failure, or (at best) as a transition between Civ4 and Civ6. Civ6 will be successful in attracting most of Civ5's fans and a good deal of Civ4's fans, though for many, Civ4 will still be the pinnacle of the series.
Of course, the whole previous paragraph is speculation. However, as someone who's monitored the business for a couple of decades now, I'm confident that it isn't
baseless speculation at all - I think I accumulated enough experience with similar situations to at least risj an educated guess. And I fully plan to return to this post in a couple of years and check whether I was right.