Let me start we few disclaimers first:

I have no insider info (from CD Project Red). All I can do is to watch carefully from the outside (as a customer) and map my observations onto my experience in the software development industry.
I realize that the gaming industry is 'special' (different from more traditional software crafting), so I'm not going to pretend I'm an expert in the field I've never directly worked in, but ... Maybe its (assumed) uniqueness is a big part of the issue?
I'm not going to guess who is 'guilty' and why - it's not about finger-pointing but about what we can actually learn from this situation. Was this avoidable? Could CDPR apply what we know from the general software craftsmanship to save the day?

OK, now I think we're all set. Buckle up.

What did happen? (from a customer's perspective)

Let's summarize the drama in a few short bullet points (for those who didn't pay attention that much):

The game (Cyberpunk 2077) has been finally released mid-December (AFAIR it has been its 3rd fixed release date, as it was postponed twice before)
Unfortunately, the product is ... well, unpolished; there are bugs, bugs & even more bugs everywhere - all kinds of defects imaginable: GFX glitches, quest problems (preventing you from progressing in the game), general instability, save-game corruption, etc.
There's a huge disproportion in quality between PS4/Xbox One and PC versions of the game (just check the difference between average rating for those platforms on Metacritic) - 86% to 54% in the very moment I'm writing down these words
Review embargo (enforced by the publisher on reviewers & media) didn't help - it has eliminated all the transparency as no one knew the real state of the game (and the expectations were skyrocketing) until the launch date
In the last few months, there was more & more (leaked) gossip about extensive crunch within CDPR - which sounded definitely like the studio had been struggling with delivery, and not everything was going according to the plans ...

What's the outcome?

Apart from all of that, the sales have been very good (it has surpassed the total costs of production within few days) - there's clearly a very good product deep there, behind all those bugs and issues. But one of the reasons for that buying rush was the studio's prior stellar reputation (of releasing top-notch products in the past) ...

And since the launch day, CDPR is under heavy fire.

Disappointed gamers rage on the Internet, legal suits are being announced, studio's reputation is damaged, their stock suffered as well (that one is probably a short-time effect, though), plenty of orders have been canceled by the players. What is more - the crunch continues, maybe even more intensely than before: CDPR needs to make the console version of the game playable ASAP to save their faces. And they still have a next-generation (PS5/XSX) versions to ship in early 2021 ...

What's different in the software release model for gaming (and why?)

Does it have too look (and end) like that in gaming? What's so special that such stories keep happening regularly (not just for CDPR but many renown game creator studios)?

games age much faster (than other software) - the longer the development takes, the more additional costs (on updating the GFX, etc.) it generates
sales are driven by hype and novelty - the majority of income happens just after the release, so the AAA games ("blockbusters" of the gaming world) have to include the "wow effect" to allure the crowds to cover the hefty up-front investment
for single-player titles, the software consumption is usually relatively short: 10h-30h (up to 100h in the most complex RPGs) but intensive; it means that you don't have time to iterate (build the game in an evolutionary way)
flawed, bad first impressions kill the immersion; you rarely have a 2nd chance (to recover something the community of players has already rejected)
game players are not only opinionated but can also be very emotional about the game (due to sentiment, expectations, etc.); they also appear quite loyal to 'brands' - the stains are hard to get off
the game release dates are determined by the season (e.g. before the winter holiday) and competition products' release dates; if you didn't make it by a week or two, you probably would have to wake few months ...
the cost (& effort) of preparing a single AA/AAA title is so high that even big studios can't afford more than 2-3 games being in the works at the same time: under such circumstances, it's hard to hedge your bets
even technically excellent product can be absolutely unplayable - the joy (of playing the game) is very ephemeral and is dependent on very volatile (& subjective) factors that can't be automatically assessed; what does it mean? e.g., you could find out that a recently completed 'feature' (game component) is not (or less than expected) 'fun' - so now you face a dilemma: what to do? Leave it (and risk players' rage) as it is or rebuild to make it better (usually by crunching)?

In practice, there are very few companies that can afford postponing their games' launches until they are 100% happy with them - like Sony's Santa Monica Studio, Naughty Dog or Rockstar Games.

Are those obstacles avoidable?

It seems that the 'core' issues are exactly the same as in other types of software:

big & long releases ('big bang' ones)
tedious, manual, repetitive QA (not really prone to any kind of automation)
risky release model (high investment up-front, success undetermined until one single moment - release)
feedback (customer perception) not known until the very end (when the final product is available)

The industry has already found some answers to address them. With the two most significant ones being:

crowd-funding
split into iterative, 'meaty' releases (to the end-user)

To be honest, both options aren't perfect:

crowd-funding still applies some pressure on the deadlines (paying customers have expectations ...), some campaigns will fail - hence increasing people distrust, game development takes years - people may not want to put their money in the freezer for such a long time
games typically are not developed in a strictly iterative way - there's an engine that powers the game, and it usually requires a lot of effort to build and polish it (it gets developed in parallel to the whole game) - early version on the unpolished version may give a false impression on game's potential (no 'wow' effect)
even disregarding the engine: game consumption time is comparable to a TV series season - imagine you'd have to wait a few weeks/months between each episode; I believe it may have killed a lot of your enthusiasm towards the product
players don't want to 'replay' content: what's the guarantee that what they've already completed won't get improved in a meaningful way in the next release? in fact - they can be quite pissed off when they have a feeling of being the "beta-testers" of the product they've paid for

There ain't no silver bullet

I don't think there's any, really.

But I still believe a lot can be achieved by improving what we already have in hand. Why is it so important? Because in fact the current model (and dramas like the Cyberpunk 2077's one) is harmful to all the involved parties: the publisher, the game developers, and the players. Yes, a patient gamer can indeed wait until the official reviews are published and make a buy/not-buy decision based on Metacritic's score - but still, is (s)he a winner by simply not buying a (flawed) game that could have been a fulfillment of her/his dreams? Only partially, IMHO.

The first step is already being made - the cultural change regarding the expectations of the purchased product. The world of gaming changes continuously - some time ago (before everyone had a connection to the Internet), patching of a game was already a big challenge - but these days the product can be developed further after the official launch. Here are some examples from the past:

the early releasing model has worked fine for some games - just to name a few: Sea of Thieves, Baldur's Gate 3, Hades (but TBH it's not true for 100% of cases - many half-baked games were pretty much abandoned by their developers for various reasons)
crowd-funding of computer games has been successful in some cases as well - e.g., people were not afraid to back Star Citizen up with tons of money (yes, it's not a typical single-player game, I realize that)
No Man's Sky has proven that the game (regardless of initial disappointment) can be successfully evolved if the honest effort is consistently applied on the way by the crew with a clear vision and skills on board
indie game development has turned up to be a reasonable alternative to the traditional game development model - still, it's more than evident that it's a complementary model, not a 1-to-1 replacement
iterative (chapter-style) release model was already successfully applied e.g. by Telltale - in their best-selling The Wolf Among Us.

A model to be

IMHO what the industry needs is a tremendous, combined success: of a VERY GOOD game with a VERY WELL executed crowd-funding + satisfactory, iterative release model (perceived unequivocally positively by the players). It's not nearly enough for it to be an excellent game. It's equally vital that it's a huge financial success that goes on par with the most successful AAA games of the past.

This will probably require some tweaking - e.g. in pricing strategy (instead of one, up-front lump sum payment - a bounded series of smaller payments that would smear the risk better among both producers and consumers). Some will oppose such changes, because that sounds like turning a product into service - there's a rich history of using such a practice as a dark pattern in sales. But TBH I think that's the only logical way out.

P.S. Regardless of the rought start, I keep my fingers crossed for the Cyberpunk 2077 ultimate success. I've read all the Gibson's books as a teenager, Cyberpunk 2020 (from R.Talsorian Games) was one of my most favourite pen&paper RPGs of all time, CDPR's Witcher 3 has delivered me one of the most immersive computer game experiences ever - this mixture is simply destined for success :)

P.P.S. Yes, I've bought the game (PC), but I decided not to play it until it's well patched and fully playable w/o glitches. There's no rush.

What can we learn from the Cyberpunk 2077 launch drama?