
I think this statistic would be more interesting if it filtered out all of the blatant cash-grab, asset-flip, AI generated shit that makes up a large portion of new releases.
Is it 19,000 releases with 10,000 actual sincere efforts at making a game, or 19,000 releases with 1,000 actual games.
And what’s the average number of reviews for actual games versus garbage?

This $1.2b in Steam sales is cumulative since around August 2020, so a period of just over 5 years. We can say $240m per year on average.
Their last annual financial statement said that they made $2.8b profit for the year ending March 2025 from their “Games and Network Services” division, so subtracting the ~$240m of Steam sales would leave about $2.56b from PS5 and PSN sales.
That’s about 90% PS to 10% Steam. It’s a nice extra market for them, but it’s not make or break.
Monster Boy and the Cursed Kingdom is top tier.

Vulnerable to going extinct.
If you read the article it briefly touches on how the “doom spiral” could affect the trajectory of a language that is not widely spoken. It’s not a great article though, it just repeats the same thing for several pages, points the finger at wikipedia instead of the content-generation farms and then fails to properly conclude the argument of their presumed hypothesis.

There are multiplayer games from 30 years ago that still have 30 people who play on the first Friday night of each month, and they will put that in their calendar and keep the game alive.
The idea that multiplayer games need huge communities of players otherwise they are “dead” is what is killing multiplayer games.
I don’t disagree. It would require manual labelling by a group of people with enough patience and understanding of gaming to be able to reliably label ~60 new games every day. I’d have thought that the Steam community was large enough to achieve this though.