back in 2019, I was thinking of getting an MBA and as part of the exploration, shadowed an MBA class at University of Washington for a day. It was so fun. One of the things they were discussing in the class that day was a case study of Valve, specifically around the Steam Machine. The team's consensus was that Valve was carefully arranging money in a barrel, lovingly soaking it in high octane gasoline, and was about to light a match.
this Steam Machine hadn’t been announced back then? Not even the steam deck, which has been a massive success.
This only goes to show how MBAs are destructively myopic.
Valve understands that inextricably tying themselves to Windows is a long-term death sentence. SteamOS represents a lifeboat for when Microsoft goes full iOS and decides to lock down Windows in exchange for taking 30% of all software purchases. Valve has been taking this threat seriously since at least 2010, which is why they've been investing in Linux gaming. Both Steam Deck and the Steam Machine are further steps toward complete independence from Microsoft.
Proton, the Steam machine, the Steam deck, etc. were probably never about making money. It's Valve's "Plan B".
They started with Proton after Microsoft suddenly made a move with the Windows store and also started bolting down Windows a bit. As with most things Microsoft that initiative quietly died over time. But at that time Valve probably couldn't afford to take any chances. It probably also made them realize they had build a castle on someone elses land.
If you are making money in the amounts Valve is, then even the simplest risk analyses is going to show that "Microsoft rug pulling you" is one of your few existential threats. Even though the probability is low or medium-ish at best, the impact is massive. Even anti-trust isn't going to save you. By the time Microsoft gets convicted, you are already dead. Just look at Netscape.