This is all true. There are no indicators that increasing revenue from more subscribers will bring net profit increases relative to the increased cost of paying licensees. This is is, without a doubt, the most expensive expenditure of GamePass. If Microsoft have twice as many users, playing twice as many games, they will - all things being equal - be paying twice as much on licensing because that is the basis on content licensing for pretty much all media. Are games different? There is no evidence to suggest that.
We don't know that this is necessarily true even for titles that have residual payments based on the number of people that play the game.
The only thing we know, based on what MS have said is that payment to developers and/or publishers is on a per title, per developer, per publisher basis. In most, but not all, cases the game creators get to choose how they will get paid for their title being on GP.
Some choose a larger upfront payment and forgo residual payments. Some opt for purely residual payments.
Residual payments could, and likely are, front loaded (larger payments for first X number of players with decreasing payments for each of the next Y chunk of players). This safeguards the developer against a situation where they want residual payments, but don't want to get paid too little if not enough people play their game to make it worthwhile to them.
Other developers have gone for a mix of a smaller upfront payment combined with smaller residual payments.
In those 2 cases, the fewer people playing the game, the more costly it is for MS, while the cost decreases as more people play it. For most developers it's a net revenue increase regardless as these are titles that are being put onto GP after their sales have started to tail off. So it doesn't matter if 10 people play the game or 10 million people play the game. The game was unlikely to gain significant sales revenue if it wasn't put onto GP. But on top of that exposure to more people gives the potential for increased revenue from retail sales of the game in addition to the revenue from GP. This may or may not happen to all developers but has happened to quite a few indie developers that have talked about it.
Because of that last point, going forward I could see a lot of developers potentially opting for lower residuals or lower overall payments if they believe that the exposure of their game to a much larger audience will likely lead to increased sales of that game or other games that they have created.
Now, all of that likely mostly applies indie and AA games. In addition to those, you can likely put AAA games that are well into vanishingly small sales territory. IE - all of these are likely have deals structured so that the developer/publisher gets are large up front payment with residuals that are either small or are reduced as the number of players increases.
Heck due to the nature of sales within the launch window, it certainly wouldn't surprise me if residuals were greatly front loaded (and large) in this case.
All of that bit is how I imagine residuals (if there are any) are structured as there are far more developers that want to be on GP than MS is currently allowing into GP on a weekly basis.
The one interesting thing that comes up is what about AAA games that enter GP either at launch or in, say, the first 6 months of release when sales are still relatively strong? Here it could make sense for MS to enter into residual payments that don't reduce over time in order to boost subscriber numbers and/or maintain current subscribers. But even here there's likely room for scaling residual payments depending on the relative "value" of the IP.
This is where the developer/publisher will have the greatest latitude in dictating terms for being on GP that would be acceptable to them. Unlike the previous group where if what the developers want and what MS determines could be profitable don't match up, they can move on to the next developer.
Obviously this means that the service would be unprofitable at the start as MS has to ensure fair compensation for developers in the case that few people play their games due to low subscriber numbers or the possibility that even with large subscriber numbers not many people play the games. Combine that with MS experimenting with a multitude of payment options for developers and losses are quite acceptable.
As the platform matures in both subscriber numbers and MS' experience with various developer payment structures, I fully expect that any residual payments will be structured to be profitable to MS and fair to the developer. I also expect that most residuals (if the developer chooses that option) will be on a sliding scale with larger upfront residuals.
This isn't like, say Netflix that still has to compete with VOD rental services or Spotify that has to give payments based on 100's of thousands (or millions) of songs being on the service (which Netflix also has to deal with). I doubt there will ever come a time when there are 10's of thousands of games on GP much less 100's of thousands. It'll likely always be a service with a curated list of games from well regarded developers or critically acclaimed games from new developers.
Regards,
SB
) 
