> ## Content Index
> Fetch the complete content index at: https://dissect.digital/llms.txt
> Use this file to discover other available public pages before exploring further.

# Xbox, Microsoft, Gaming, and AI: Part 2 - AI is a very safe bubble for Microsoft
- URL: https://dissect.digital/xbox-microsoft-gaming-and-ai-part-2-ai-is-a-very-safe-bubble-for-microsoft/
- Published: 2026-09-23T06:33:50.000Z
- Updated: 2026-09-23T06:33:50.000Z
- Author: Digit

Blue chip, larger industry, flexibility of hardware, and mitigating risk  
  
Businesses are generally interested in this thing called "making money" in order to maintain something called "financial liquidity," and where executives seek to "[exploit](https://arstechnica.com/gaming/2008/11/activision-if-we-cant-run-a-game-into-the-ground-we-dont-want-it/)" the intellectual property of the company, and thus, by extension, the many workers at the company.  
  
Being slow to write this article, Jason Schreier has recently come out with [their own piece talking about how video games have become a mature industry](https://www.youtube.com/watch?v=%5FHJXSt7ofOU&pp=ygUOamFzb24gc2NocmVpZXI%3D ) rather than a growth market. With investment declining in recent years as we pull out of the primary COVID-era isolation and digital life, and the bit of a bubble that was created in gaming crashing (and all the nonsense like NFTs that got glued to them and such), the industry is in a state of contraction. For more about the contraction, check out [the data from Matthew Ball](https://www.matthewball.co/all/stateofvideogaming2025) (who you may recognize as a current exec at Xbox today), and for additional context and analysis, [this video from Alanah Pierce](https://www.youtube.com/watch?v=9HM9nmqNioQ&t=180s) about the report. Additionally, you can check out [Amir Savtvat's presentation on the state of the industry as posted on GamesIndustry.biz](https://www.gamesindustry.biz/the-big-picture-what-you-need-to-know-about-the-ongoing-games-industry-reset). For those who may be unfamiliar, Amir spent much of his time over the beginning years of COVID amid the crash trying to reconnect game developers who were laid off to new jobs, and has been a savior of countless people's careers due to his aid in this area.  
  
The point is, there's not much more money to try to pump out of people.   
  
Just as any mature company and industry, we would call Xbox(-Activision-Blizzard-King-etc), a blue chip company that is stable and not going to grow much. This means monetization is going to be poor because there's not much new area to tread, much as Jason points out about PlayStation and digital games in his video. The only place to go is exploitation.   
  
Here's where I want to take a second in the structure of this and say there are two paths. One where Microsoft could support their games, and one where they aim for AI and data centers. First, I'll talk about how they could be a good game publisher, and then follow that up with the reality they chose.   
  
Now I would argue that if Xbox/Microsoft really wanted to make money, there is a good way to do so that is not so exploitative. To call back to Jason Schreier again, he has explained why [studios love the risk of the live-service market](https://www.youtube.com/watch?v=rBwuxx5XSYc&pp=ygUhamFzb24gc2NocmVpZXIgbGl2ZSBzZXJ2aWNlIGdhbWVz), and [how indie games are 1000:1 odds of getting a viral hit](https://www.youtube.com/watch?v=ZKND9udOoTQ). While Jason doesn't say it's impossible, he mentions how the industry executives do not seem to have the stomach for it, which is rather ironic. I would argue that indies at scale might have more potential because a developer or publisher can develop a following or a cloud over their head. With some aid, money, and pixie dust, there's certainly an area of growth that could make more money if they really wanted. One can look at developers like Strange Scaffold with their medium-sized cult following, or Mintrocket, who came out with Dave the Diver and is still raking in money and expanding on the game with new free and paid DLCs. A place I'd like to write about later is how Kawloon Nights chooses every single correct horse to fund, and by understanding what makes a fun game, has helped numerous successes as a group that simply funds games they think have promise. If Microsoft had the follow-through to take care of the studios they acquired and organized more like how Mintrocket used to be an incubator for Nexon before going independent as a subsidiary less tied to the central company. As much as I love Hellblade, it is ridiculous to part ways with so many large games right after announcing them and when they are quite far in production. In the "indie" conversation, though, there's no reason they had to be huge bets in the tens or hundreds of millions of dollars.   
  
Some may want quality games but also may be willing to take on failures like [Playstack (godspeed under their new owner, ICM), who says it would be a mistake to have a 100% success rate and that 70-85% is brilliant](https://www.youtube.com/watch?v=-mYLopeLFHQ) because otherwise they're not serving the indie space properly. Part of that interview also mentioned how 'fearless' indie games can be and that they take on lower-budget games more often because they do represent a chance of success with a lower risk that makes it work for them.   
  
One only needs to look at Dave the Diver, or Peak, or REPO, Minecraft's origins, Shift at Midnight, Big Walk, Untitled Goose Game, Abiotic Factor, Look Outside, Consume Me, Celeste, Hollow Knight and Silk Song, YapYap, Megabonk, like any Hazelight game, Ball x Pit, Disco Elysium, Pico Park, Sticky Business, Beat Saber, Stardew Valley, etc etc etc. You get the point.   
  
Funding these games doesn't even have to produce polished results as many of the most profitable games are "low quality" user-generated content fueled by microtransactions, because in reality, many players just don't care about graphics at the end of the day and are more than willing to put up with some jank, whether it's a kid who just accepts it or an adult who has seen that just because you can see every pore on a face doesn't make it a good game. We can even look at others. We also see an inefficiency in studios' current design. Right now, games that used to be able to take up less than 10GB are taking up well over 100\. Marvel Rivals takes up 124GB of space despite being a stylized game where 4K resolution textures are never going to be noticed because it's a third-person game (even uninstalling the high-res textures that they finally allow leaves it taking up 64GB at the time of writing). Arc Raiders was smart about this. They had some good seed money, but they themself have admitted to the fact that a first-person camera would not look good due to poor close-up detail. Despite this fact, players rave over the visual style and how good it looks because of the visual/artistic direction.   
  
When you make games affordable, when you let a team work well and have a vision, it's incredibly possible to make an outsized amount of money on a project. Many will likely flop. I don't think Trine is doing great these days, or that all that many people are aware of Tiana Sisters, but when you can set up a team of a dozen or so people with the possibility to make $10+ million, I would wager that the game is smaller than people think when, and this is important, **you have established positive brand cachet**. This is what Xbox could have been, and people loved many of the things they were showing. Given another even year or two, the numbers could look very different and more profitable. Notably, you don't make money before it comes out. I think one only needs to look as far as Annapurna (even if not too low budget) to see that.   
  
The other reason I believe that this is possible is the current market and what gamers are looking for. Sure, there are people looking for their "everything game" like Grand Theft Auto or Fortnite, where they will spend all their time, but many aren't. Most of the people I know bounce from game to game. As adults with disposable income, there are loads of people who are more than willing to swap their morning coffee budget for a game to spend a few hours enjoying themselves with friends, or who may not find the $3-10ish dollars all that significant in the first place. Additionally, those of us who aren't chasing down nostalgia to get disappointed by Halo: Campaign Evolved, or playing 100hr sagas of games, have a lot more time and are looking for new and interesting twists on what we love to do. When there is another half of a market that isn't a 13-year-old looking for the single game they're getting for the next 6 months, you start to see the indie renaissance and how small independent studios/creators are making millions off of twisting even singular mechanics in a game and giving something interesting to think about. The one thing that really connects a lot of the most popular games we see now is that they are co-op. Roblox, Fortnite, Split Fiction, It Takes Two, Peak, etc are all social experiences. Once you get someone to enjoy it, there is an incentive to get others to play it too, and many people look for an excuse to spend time together rather than just playing a game. For more on the idea of social gaming, check out [Razbuten's recent video](https://www.youtube.com/watch?v=knXjtvIoBpg) on how they find value in these experiences. Now that's all my take, but let's get back to reality.   
  
**\[Note: As of a few days before this little note on Sept 4, 2026, it seems Double Fine is going to try making quick-turnaround indie games using their new little label "Double Fine Action Labs" and announced their first game supporting this effort,** [**Thank You Bus Driver**](https://www.doublefine.com/news/announcing-thank-you-bus-driver)**. DF tells readers to learn more about their recent efforts in a** [**Games Beat interview linked here**](https://gamesbeat.com/how-double-fine-is-charting-its-destiny-outside-of-xbox-starting-with-a-kickstarter-campaign-tim-schafer-interview/)**. A big shout-out to Double Fine for being brave enough to try this approach, and best of luck to them. If you want to support their current experimentation, you can also support their somewhat-annual Amnesia Fortnight game jam and creation at the Kickstarter** [**here**](https://www.youtube.com/redirect?event=video%5Fdescription&redir%5Ftoken=QUM4Zm9rVFBfN0p0VDJmSkxPSzhDOGVBcFJ2SXxBTl9pYzRjemdJQUFMbm8yQlZvVV9KQjFpTkZOMjVKWWI2aEZpWmtCbnF6b2R4N2VjMlFXSW4tS0Q3UnRtWlczdkRQRzZmTlRBOTh4Y3p5dnNVX25hNldzVEJEbWZxVWNscWQx&q=https%3A%2F%2Fwww.kickstarter.com%2Fprojects%2Fdoublefine%2Famnesia-fortnight-2026-indie-reboot&v=3Bk%5FAFXv3VU)**.\]**  
  
So back to the AI-shaped monstrosity in the room, why is Microsoft pushing so hard in this direction? We know they care about profit, especially for high-growth, short-term profit. Most people would see this as high risk, but I would argue that it isn't. I think this is an extremely safe move compared to other companies doing the same thing. If Xbox is gutted, then they aren't going to lose much money, whatever happens on that front. It also frees up cash flow to go towards AI. But the real piece of the puzzle is **that** Microsoft is in cloud computing. They have Microsoft Azure**,** which is their server infrastructure that they can use for any number of things and rent out capacity by way of units of "compute hours" or tokens. They might spend money on electricity training the models, but they own the computers doing the work, so it's cheaper. If someone else were to make the same AI model, they'd have to rent the infrastructure and pay the utilities. Then even if AI is a bust, they still own the servers. Just like so much runs on Amazon Web Servers/Services, same for Microsoft Azure. This may be familiar to you if you've seen [People Make Games' reporting on how the Israeli Military used Microsoft web infrastructure to monitor all communications in the illegally occupied country of Palestine](https://www.youtube.com/watch?v=Ybj10537yi4). The cost for Microsoft is in buying hardware and building out data centers. After that, they have a vastly increased capacity to offer software as a service and make recurring revenue from anything from website hosting to cloud storage to government infrastructure. Right now, the AI boom is just giving them a payload of cash to pursue this handsomely profitable future that will pay dividends in the long run. I think a fair number of people realize LLMs are nowhere close to the future (and relatively unprofitable, as companies in China are putting out better ones with open weights for free), but if they can get another billion dollars to build out 1000 more servers, then they have secured expansion funds and have a brighter tomorrow. During the current manufactured AI boom/bubble, they can also recoup these costs quickly, as every competitor needs to be careful not to spend themself out of business. This is the biggest gamble depending on the scale of the numbers and internal balances. Yes, the "I want to be the next '.com' or 'iPhone'" thing can still exist, but it's not where the money is.   
  
(I believe there may also be an element of induced demand here, as those that choose to use AI and become more dependent on it will continue in the future, thus providing some level of stable core customer base as said customers want convenience over correctness or learning how to use the tools at their disposal.)  
  
Where the money is, whatever it will be, is in software. We are seeing a relatively stagnant hardware market and not enough capacity to make what we need already for our home computers. Every revolutionary thing is in software right now, from useful things like [AI upscaling in games](https://www.youtube.com/watch?v=nzomNQaPFSk) to make your expensive GPU last a couple more years, machine vision for accessibility and industrial tasks, [machine learning in biology and pharmaceutical applications](https://www.youtube.com/watch?v=P%5FfHJIYENdI), AI body tracking for body estimation motion capture, movie rendering pipelines, innovations in software like Unreal Engine's [Nanite](https://dev.epicgames.com/documentation/unreal-engine/nanite-virtualized-geometry-in-unreal-engine?lang=en-US) or [Lumen](https://dev.epicgames.com/documentation/unreal-engine/lumen-global-illumination-and-reflections-in-unreal-engine?lang=en-US) and Nvidia's [Mega Geometry](https://developer.nvidia.com/blog/nvidia-rtx-mega-geometry-now-available-with-new-vulkan-samples/) and [neural texture compression](https://research.nvidia.com/labs/rtr/neural%5Ftexture%5Fcompression/). Who knows what the next gold rush is? It doesn't matter if you know roughly where it will be and you're selling the shovels.  
  
This is an even safer bet when you run gaming servers as Microsoft does. All that Xbox Cloud Gaming compute power that runs GPUs being able to be flashed with software and given a new purpose, depending on whatever the client at the time wants? That's convenient. Xbox somehow regains ground and needs to divert processing power back to running servers for games and doing real-time flight tracking and simulation for Microsoft Flight Simulator? Sure, you can do that too. Some tasks are going to be better on CPUs vs. GPUs, but at the end of the day, it'll generally be able to work in some way and do so with varying levels of efficiency that are better than paying someone else for it, and you can make money off of doing it for others. Any new startup can't compete in terms of cost, and their new LLM could be obsolete in a day. Microsoft has a moat and a solid foundation. I would even go so far as to hazard a bet that this server capacity-based argument lines up well with how [Xbox cloud gaming services are putting time limits on all the users with a maximum tier of 15 hours a month](https://news.xbox.com/en-us/2026/09/03/xbox-cloud-gaming-changes/). Not even enough for an hour a day, at [a price of $23/month](https://www.xbox.com/en-US/xbox-game-pass#join) at the time of writing. A subscriber doesn't get even a dollar an hour on the shared server hardware and, at the end of the year, would be paying \~$300 with tax included. That places the most dedicated players at a casual level of engagement, but paying enough that it is *well* worth investing in their own computers and freeing up Microsoft's servers anyway. While the average gamer only plays a couple of games a year and the change to subscriptions will impact only an estimated 4% of people, it does beg the question: why bother unless there is a bigger fish you can make much more money off of with that capacity? The decision is slightly confounding, but we do see that positive brand cache does not matter to corporations. Maybe there is another strategy, and that is at play, but even at 4% of the [roughly 30 million current Game Pass subscribers](https://www.gamesindustry.biz/xbox-game-pass-reportedly-boasts-around-30m-subscribers), that's still 1.2 million players that servers must be reserved for. \[Note: It seems after backlash, Xbox has "clarified" that existing users will be grandfathered into unlimited hours **depending on the user's region**, though it seems the US and some others will not be saved. Check out [Skillup's video](https://www.youtube.com/watch?v=YSE6JEI-DfE&list=PLbcyJSeTLtlzGRENniEdwxC2C6mXOXlcz&index=2) at [18:06](https://youtu.be/YSE6JEI-DfE?si=TJ7Ob3mUI6dbLiPS&t=1086) for a brief update on some of the details, and link to their [written coverage here](https://thisweekinvideogames.com/news/xbox-introduces-cloud-streaming-limits-for-game-pass-purchase-options-for-ad-hoc-use/).\]  
  
As far as AI companies go, Microsoft is as safe as can be. The real question is what will happen to Xbox specifically?