I don't know that this is true. At a certain point with enough subscribers it can become sustainable. If Microsoft were to be able to get 30 million people to subscribe to GPU at current full price, which is admittedly a large lift, that's $450 million in revenue per month. That more than covers the cost of releasing one AAA title per month with hundreds of millions left over. Delivering just 3-4 AAA first party games per year would make the yearly subscription cost well worth it for consumers.
That will probably never happen though because no service has an ARPU in line with its MSRP on a per-user basis. For example, PS+ had an annual ARPU of $52.08 in 2021 (or FY 2021). However that was with a sub count of 48 million.
I've figured that GamePass's likely annual ARPU is too low at the 25 million subscriber mark to where 30 million subscribers would make a difference. They would need closer to 72.6 million subscribers to predictably generate $450 million per month, or $5.4 billion per year in GamePass revenue. And, given other sub services like PS+ and NSO have shown to have less than 50% install base saturation, if MS were to try getting those numbers through just console, they'd need ~ 145 million Xbox owners, which is simply not happening (especially if, even if you could count XBO towards it, that would be a dead platform commercially and vast majority of its owners would have moved on to Series S & X Xbox console-wise).
Although MS have said they expect future GamePass growth to come from outside of console. That said, there's no telling how much ARPU subscribers on PC & mobile would be paying, if it would be close to what subscribers on Xbox are paying (on average) or much lower. Because in the case of the latter that would just increase the amount of raw subs they'd need to hit a steady $450 million per month in service revenue.
If Microsoft could get their act together and drop a big game every other month it makes a lot of sense at the right number of subscribers to be able to keep the recurring revenue flowing. At scale it works better than the one time purchase model. That's why most major software you buy these days is on a subscription basis instead of a higher cost perpetual license. Surely people here are going to continue to hold on to the idea that somehow video games are different than pretty much every other form of digital entertainment and can't thrive on a subscription model. But as an alternative delivery method Microsoft is betting they can.
The kicker, though, is that "major software" you're referring to is generally
NOT entertainment software. It's productivity & business software. Programs for clients who use them for work and to help them in turn have that work generate revenue for them as per their jobs/careers.
Entertainment software, such as games, don't have that type of value proposition. That's why subscribing to MS Office or Adobe Photoshop is an easier sell than subscribing to a Netflix-style gaming service. Maximizing your value out of MS Office or Photoshop just means developing a skill set to pursue work that requires the software, creating content and then generating money off of that content, which is going to be a requirement anyway if your profession requires it.
To get similar maximized value out a Netflix-style gaming service, you have to play a
LOT of games and try doing so through as much of the catalogue as possible. However, most people tend to commit to just a very small selection of games, or even just a single big game release, at a time. Very few people in general are playing say 10 games a week, in any meaningful capacity. And quick-sampling a game for a few minutes is not the same as actually "playing" it, in terms of making deep progress and getting invested into the story, mechanics, characters, levels/world and such.
In general, most customers only play a handful of games per year anyway, and quite a few wait for sales to kick in to buy what they want. So the only way most any subscription service can beat that in value is to have ways where a subscriber can pay for the service at a much lower price than the typical cost, at which point they're severely lowering the ARPU meaning you need an exponential amount of additional subscribers to make up for it.
The trick is to get enough devices into homes to build the install base to drive subscriptions. That's why Microsoft is selling the Series S for so cheap this holiday and letting other companies offer it as an incentive for things like internet and mobile phone services. They built Series S to drive this subscription model because they're betting that a lot of people are going to prefer to consume games like they consume movies and music.
Except you
CAN'T consume games like you can music or even movies without simply sampling the games. Even most of the average indie games take more time to play through than a big Hollywood 2.5 hour blockbuster. Let alone the bigger AAA games, that require hours of investment to really start harvesting everything they can offer. No matter how cheap a service can be offered for, or the games for that matter, time is priceless and you can't buy more of it.
If the idea is that people will want to get their games digitally, then they don't need a service for that. If the idea is that they want them digitally & cheaply, a service
COULD provide that but at the expense of having lower ARPU and needing to appeal to enough additional customers to balance that ARPU drop out. For GamePass specifically, at the rate it's been growing, it would not have enough subscriber growth to generate enough revenue wherein the budgets for several AAA games can be covered, for several years. If ever.
Granted, since the games aren't solely being provided through the service, the traditional sales model still exists, so the service only needs to account with enough revenue to cover a portion of the budget otherwise. However, we still don't know the extent to which Day 1 inclusion of big AAA games in GamePass has on game sales. What we can infer is that, it doesn't lead to a boost in sales and in fact might reduce them. The fact games like Halo Infinite and Forza Horizon 5, both AAA releases, haven't charted NPD outside of their debut months speaks volumes. And while Halo is generally considered a mediocre game with terrible content release timings, Forza is said to be a great arcade racer, in general, so lack of quality can't be a reason for its lack of big sales unless now it is a reason.
That's all important to consider because for any lost sales, the service itself has to pick up the slack in revenue, but there are so many ways to get GamePass on the cheap (or free, even), that as long as those known loopholes continue to exist, MS could be in a situation where the service can't generate enough revenue to ensure AAA software budgets (keep in mind not 100% of the revenue would even go towards software budgets; some of that is going to be stashed as reserve, a lot used to pay for expenses, licensing rights, etc.), yet the games arne't seeing enough direct sales to generate the needed revenue that way, either.
It's a potentially very troublesome scenario.