The Steam Machine is quietly dying

Valve has a lot more market power than Nvidia when it comes to this. They could e.g. make it part of their standard contract going forward. Or even if they don’t make it mandatory, they could offer extra promotion slots on Steam to streamable games. The latter kind of indirect subsidy has worked for them before, the number of games with native Linux support is kind of crazy in comparison with the market share.

In fact, a cloud streaming service would neatly explain they’re still continuing to invest into Proton (and Linux gaming more generally). At this point it can’t be just be insurance against Microsoft closing up Windows.

The fact that their Steamboxes ran on Linux is one of the big reasons why they failed, same with Stadia. If they’re counting on Linux gaming to help power their way into cloud gaming, they are doomed. Doubly so because they seem unlikely to even go to the somewhat muted lengths Google went to get people to port their games.

But yes, I agree that Steam has massive scale, but like Nvidia they have a vested interest in keeping publishers happy. Would they mount a lawsuit to force companies to let people play their Steam library in the cloud? Hard to say. I guess I think it’s marginally more likely than Nvidia because Gabe is a purist a bit like Tim Sweeney.

Using Windows for cloud gaming is doomed at scale, due to both the licensing costs and the licensing uncertainty. Microsoft has already done some brutally anti-competitive things around Windows licensing to win in the public cloud market. They’ll do the same the moment game streaming becomes a big business on any competitors stupid enough to build around that.

Steamboxes failed because they were far more expensive than consoles, and with access to a much smaller games library than an equally priced computer would have.

That doesn’t apply to the streaming market. Amazon Luna lets you stream games from a umm… 75 game library. Xbox and Stadia have ~200 games. PS Now has 700 games, but they’re mostly obsolete trash. GeForce Now has 1k games. Steam has 6k Linux native games, and 6k which Proton is claimed to run perfectly.

Linux won’t be the limiting factor around getting a competitive library.

No, of course not. Lawsuits would be silly: they take ages to resolve, and the outcomes would be uncertain. Instead they’ll leverage their near-monopoly around PC game sales into this new market. Maybe by force, maybe they’ll be a bit gentler and use non-monetary incentives. Hell, maybe they don’t even need to do get their hands dirty, and can just count on gamers to review bomb games that opt out.

We’re not talking about hitting a button to make a Linux port. Developers can barely get their games out on the mainstream platforms in time. Telling them they have to also make a Linux version isn’t going to win many friends.

Sure, and odds are that this is why they’ve been investing in Proton. They certainly haven’t been doing that for the Linux desktop users.

I guess I had forgotten about Proton, or maybe never really knew how it worked, but hey, that’s cool. According to this page, there are over 14k Steam games that already work on Linux through Proton, which is a pretty huge starting library if they launch a cloud gaming solution.

Maybe that’ll be the difference maker, but again, the tech part is hard, but the fact that some publishers aren’t super keen about cloud gaming in general is another hurdle.

If they can launch it at $300 we’ll talk. But I have my doubts that Valve can make this make sense at scale.

It would be neat if it ran android with a cheap SoC and was intended as a Geforce Now streamer device.

Whatever it is, if it is, I doubt it would be locked down and you couldn’t run a browser, and thus GFN.

That seems the most likely scenario to me, for sure.

But why would Valve make it and not Nvidia? No, I think this is supposed to play games directly, which is one reason why it’s dumb.

My thought is it’s a collaboration between Valve and Nvidia to stream Steam games via the GFN service. All you need is a 7" 16:9 1080p screen with half a gamepad on either end and a cheap phone SoC to stream games. And it’ll run Android so you can use it for xCloud or watching Netflix or whatever else you want too. Could easily cost $200.

But knowing Valve it’ll probably be an $500 device designed to compete with the Switch running Linux (not Android) and nobody will ever buy one.

Well, the Ars story says it’s most likely AMD or Intel providing the chips, not Nvidia.

My guess is $799. Valve is terrible at making mass market hardware, but luckily they don’t care about making money. If this is like the Index they can cater to a very small group of hobbyists.

That is the rumor, yes. If it’s streaming games, if doesn’t matter who makes the chips. Performance doesn’t matter, you just need to decode 1080p video, which any given microwave oven can do these days.

My bet is actually on $799, running a full copy of Windows 10 with Steam big picture auto-loading as the shell, 16GB RAM, semicustom quadcore Zen3 mobile CPU with an integrated 24 CU RDNA2 mobile GPU, similar to the XSS. 7" 16:9 1080p screen, 512GB NVMe SSD. Weighs 2 pounds. Battery life 3 hours gaming, 15 hours playing Netflix.

I just can’t imagine Valve trying to push Linux to consumers again, even with Proton. Too far-fetched.

My guess is that they’ll launch this together with their own game streaming service. The only thing that’ll run locally is the Steam Big Picture UI. In fact, they won’t even make it possible to install any apps locally. Anyone wanting to use other streaming services with this would have to use the browser.

The idea of it running games locally just doesn’t make sense to me. A lot of people are already making those portable gaming PCs. They’re pretty awful looking experiences, and far too expensive to ever become popular. Anyone buying one of those will likely be using it mostly for Steam anyway. There’s just no need for Valve to make one more of those.

But a streaming-only handheld? There seems to be nothing in that market yet from third parties, it ties together nicely with the game streaming work we know they’re working on, and I bet they could hit a really attractive price point and battery life with that.

I don’t think they’d collaborate with Nvidia on this. We know that Valve is working on game streaming of their own, so Nvidia is a competitor. No point in boosting them up. The UX for GFN is absolute trash, you can’t sell that in a consumer device. Also, it’d be Valve selling the low-margin hardware and Nvidia raking in the sweet subscription revenue. Doesn’t feel like a great deal.

IIRC the Steam Link was pretty aggressively priced at launch. It’s just that it wasn’t very compelling as a product, there were already good ways of playing PC games on a TV. But a Switch form factor streaming device? If it allows both local and cloud streaming, it’d probably make me toss the Switch permanently in the junk drawer.

I think you’ve invented a compelling product. One issue with a handheld streaming device is what happens when you’re away from home? It’s a brick unless it uses cellular and that opens up a big can of worms with data caps.

I’d make it wifi-only, if somebody wants to play over a cellular connection then they’ll have to tether to the phone. >95% of my Switch usage is in places with good WiFi anyway, even if not at home. (“Everyone generalizes from one example; well, at least I do”).

And no point in wasting their BOM on integrated cellular, nobody wants to mess with extra SIMs and I think for something like this they’d want to get the price as low as they possibly can. (Would $99 for 1080p @ 6 inches with non-detachable controllers be doable?)

If it is a streaming device, I don’t really see the upside in this, for the consumer or Valve, versus just buying a Razer Kishi. And if it isn’t, I’m not sure why it makes any more sense than the others out there like the Aya Neo, which barely makes sense as it is.