The case against subscription cloud catalogs

If you've never tried cloud gaming, the easiest pitch in the industry is Game Pass Ultimate. One price, a few hundred games, day-one Microsoft first-party releases, the whole thing streams to a Chromebook or a Samsung TV. There's a reason it's the most subscribed-to cloud service: it removes the two biggest sources of friction (which games do I own? what hardware do I need?) at the same time.

PS Plus Premium does the same trick for Sony's catalogue. Luna does it lazily but cheaply via Prime. The shape is the same: don't worry about what you own — we'll lease you a library, and we'll stream it for you.

Game Pass rotates titles out roughly every quarter. PS Plus Premium does it less aggressively but it does it. Luna shuffles its Prime channel monthly.

If you played a game on a subscription catalogue and didn't finish it, you have two options when it leaves: lose access, or buy it again. The second option is the lie at the centre of the catalog model — the games aren't really part of your subscription, they're rentals with a finite end date you don't get to pick.

This is fine for games you play for ten hours and shelve. It is not fine for the games people care about most: 200-hour RPGs, the MMO they've raided in for three years, the racing sim they've spent eight months tuning a setup for.

GeForce Now's model is the inverse: you own the games on Steam, Epic or Xbox PC, and NVIDIA streams them to you. If NVIDIA loses a publisher deal tomorrow, your saves still exist on Steam, your time invested still exists, and your only loss is the ability to stream that one title — not the title itself.

Boosteroid works the same way. So does Shadow PC (which we don't cover because it's a desktop-rental product, not a games product). The throughline: your library is decoupled from the streaming infrastructure. That's the structural advantage.

Subscription cloud catalogs sell you optionality: try anything, finish nothing, move on. Bring-your-own services sell you continuity: this is your library, we'll stream it to whatever device you're at.

If you're 17 and you want to dip in and out of forty games a year, the catalog model is genuinely better. If you're 35 and you have three games you've played for 200 hours each, you should not be on Game Pass for cloud. You should be on GeForce Now, streaming the things you've already paid for.

Game Pass Ultimate is $19.99/month. That gets you cloud streaming plus a console catalog plus EA Play plus day-one Microsoft releases. GeForce Now Ultimate is $19.99/month and gets you streaming alone — you also need to pay for the games.

On a strict dollars-per-hour basis, Game Pass usually wins for low-volume players. The cost compounds in the other direction over time though: a Game Pass subscriber who plays 30 hours/month for five years has paid $1,200 and owns nothing. A GeForce Now subscriber who plays the same hours has paid the same in subscription, plus maybe $400-600 in games — and owns those games forever, including outside the cloud.

We give Game Pass and PS Plus Premium good reviews. They are, on the metrics that matter to most casual players, excellent products. But we'd rather see the cloud gaming industry move toward bring-your-own as the default and curated catalogs as the supplement, not the other way around.

The structural risk for catalog services isn't the streaming infrastructure or the price — it's that they're selling something they can take back. That's a worse foundation for a long-term consumer relationship than 'this is your library, we'll stream it' has ever been.