"Unlimited" is doing a lot of work in tech marketing these days, and it never means what you think it means. Unlimited data has throttling tiers. Unlimited storage has fair-use policies nobody reads until they hit them. And as of today's Xbox Wire post, unlimited Xbox Cloud Gaming has an expiration date, and depending which Game Pass tier you're on, the promise it's breaking is either a year old or six years old. Because this isn't just Microsoft trimming a feature it added last year to sweeten the newer, cheaper tiers. Cloud gaming has been a Game Pass Ultimate perk since the service launched in September 2020 — it's one of the two or three things that has always justified paying the most for the top plan, alongside day-one releases and the EA Play bundle. Unlimited cloud streaming wasn't a bonus tacked onto Ultimate at some point along the way. It was part of the pitch from day one, for six straight years, before today. If you've been an Ultimate subscriber this whole time specifically because "play anywhere, no console needed, no cap" was baked into what you were paying for, this is the tier where that promise just quietly expired — not the newer, cheaper tiers where you might reasonably expect some corners cut. Starting in November, Xbox Cloud Gaming stops being an unlimited perk of your Game Pass subscription and becomes a metered one: Game Pass Ultimate: 15 hours a month Game Pass Premium: 10 hours a month Game Pass Essential: 5 hours a month Run out, and you'll be buying additional cloud playtime through the Xbox Store. Microsoft hasn't said what that costs yet — pricing, packaging, and whether unused hours roll over are all "coming in November," which is corporate-speak for we haven't finished deciding how much you're going to hate this. The Justification, Verbatim Microsoft's own framing, pulled straight from the announcement: "We're introducing this model because the cost of providing cloud gaming grows as more people use it and play for longer. Moving to monthly limits allows us to keep offering the service while continuing to invest in its reliability and performance. We understand that for some players the practical result is a higher cost." That's a remarkably honest sentence for a corporate blog post, and I mean that sincerely — most of what I write about in this series is companies burying the actual admission under three paragraphs of mission-statement filler. This one just says it: costs went up, so you're paying more, and yes, we know. Credit where it's due. It doesn't make the change good for you, but at least nobody's pretending 15 hours a month is a gift. "Only 4% of Subscribers" Is Doing a Lot of Quiet Work Xbox says this change will affect roughly 4% of Game Pass subscribers — the framing is meant to sound reassuring, like don't worry, almost nobody uses enough cloud gaming for this to matter. Flip that statistic around, though, and it says something less flattering about the product: if only 4% of subscribers are streaming more than 15 hours a month, cloud gaming was never the mass-adoption feature Microsoft spent the last year selling it as when it rolled unlimited access out to Premium and Essential tiers. It was a marquee bullet point on a pricing page that most people were never going to use enough to cost Microsoft real money — until, apparently, it did. And for context on just how stingy 15 hours actually is: Nvidia's top GeForce NOW tier includes 100 hours a month. Microsoft isn't just introducing a limit, it's introducing one that's roughly a seventh of what its most direct competitor already offers. The Fortnite Tell Here's the detail that actually explains the "why" better than the blog post does. During the FTC's trial over the Activision Blizzard acquisition, former Xbox President Sarah Bond testified that Microsoft lost money every single time someone streamed Fortnite through the cloud. That's not a one-off outlier — Fortnite is one of the most-played games on the planet, which means Xbox has been running its flagship cloud feature at a structural loss on some of its highest-traffic content for years, and just now decided to do something about it. That's the real story under "costs grow as more people use it." It's not a hypothetical scaling problem, it's a specific, already-testified-to-under-oath problem, and the fix is what it usually is in this series: shrink the thing everyone was told was unlimited, and sell the difference back a la carte. The Steam Deck Problem Microsoft Has Never Bothered to Fix Here's who I think actually eats this change, and it's not a hypothetical: anyone playing Xbox games on a Steam Deck, or a Windows handheld, or a phone, because they don't own an Xbox at all. For a lot of people, cloud gaming was the Xbox experience — not a bonus feature layered on top of a console they already had, but the entire reason Game Pass was worth paying for on hardware Microsoft never shipped a real client for. And that's not a small caveat. There is still no official Xbox app for SteamOS or Linux. What you actually get is a website — xbox.com/play works in basically any modern browser, Chrome, Chromium, Vivaldi, Brave, probably Firefox too, not just Edge, whatever Microsoft's own documentation would have you believe. (Edge gets top billing in Microsoft's own guides for the obvious reason that Microsoft owns Edge, not because the other browsers don't work.) People dress the page up to feel native with kiosk-mode launch flags or lean on unofficial clients like Better xCloud or Greenlight to claw back features Microsoft never bothered building in, like proper mouse and keyboard support or finer bitrate control. Years into Xbox Cloud Gaming being a marquee subscription feature, the closest thing to an official on-ramp for the single most popular gaming handheld on the market is still "here's a webpage, and by the way, use our browser." Now think about what those sessions actually looked like. A lot of that Deck cloud time wasn't people slowly working through a 40-hour RPG — half of those run fine on Proton anyway, so there's no reason to burn cloud hours on something Steam can just install and run locally. Anti-cheat support on Linux has gotten better too, but it's uneven — plenty of studios have flipped the switch, but it's disproportionately live-service shooters and competitive MMOs where publishers still hold the line, often citing Linux/SteamOS as the scapegoat in the broader fight against cheaters, whether or not that's actually where most of the cheating comes from. What's left is the narrower, genuinely awkward case: games that either aren't on Steam at all — Game Pass exclusives, PC app-only titles — or whose anti-cheat still hasn't been toggled on for Linux, concentrated in exactly the genres people play in short bursts. Fortnite spent years in that second bucket before Epic finally enabled it. That's the stuff people were actually streaming, because it's the format cloud gaming is good for and the only option Proton couldn't cover. Which loops back to the Sarah Bond testimony above: if Fortnite specifically was losing Microsoft money on every cloud stream, and Deck owners with no other way to play Xbox games were disproportionately the ones streaming exactly that kind of game, this cap isn't landing on some diffuse 4% of power users. It's landing squarely on the people who were using cloud gaming exactly as intended, on hardware Microsoft was happy to court but never quite committed to supporting properly. The Part That's Genuinely a Little Interesting Buried in the same announcement is something that cuts the other way: Microsoft is opening Xbox Cloud Gaming to people who don't have a Game Pass subscription at all. Starting in November, you'll be able to buy cloud playtime hours directly through the Xbox Store and stream games you already own — no recurring subscription required. That's a genuinely useful option for someone who owns a handful of digital Xbox games and just wants occasional cloud access without paying $22.99 a month for a subscription they'd barely touch. It's also, not coincidentally, the same mechanism that makes the Ultimate/Premium/Essential limits digestible from Microsoft's side — the same hourly-purchase pipe now serves both "casual non-subscriber" and "Ultimate subscriber who blew through their 15 hours by the 12th of the month." One new billing surface, two audiences to sell it to. My Honest Read Strip away the framing and here's what I actually think happened. Cloud gaming was supposed to be the thing that made Game Pass sell itself — no console required, play anywhere, the subscription is the platform. That's the entire premise behind Microsoft's own "This is an Xbox" ad campaign, the one that slapped the logo on a fridge, a toaster, a laptop, basically anything with a screen and Wi-Fi, to make the point that hardware didn't matter anymore. That was the pitch when Ultimate got cloud access in 2020, and it was still the pitch a year ago when Microsoft handed it to Premium and Essential subscribers as the headline addition. The bet was that removing the hardware barrier would blow the subscriber numbers wide open. It didn't work, or at least not enough to justify what it cost. In the same twelve months Microsoft was making that bet, they also hiked Ultimate 50% and watched subscribers leave in numbers big enough to force a partial walkback six months later — which tells you the growth story wasn't holding up even before you get to the server bill. And Xbox is still losing money on the whole operation, by its own CEO's admission, badly enough to justify cutting 3,200 jobs and divesting five studios this year. Worth noting: Microsoft already quietly scrubbed "This is an Xbox" from its own marketing earlier this year, pulling the campaign page and most references to it around the same time it announced a new console-PC hybrid — i.e., right around when the company started walking the message back toward "actually, buy the hardware." Today's cloud gaming caps are the second half of that reversal. First they backed off the marketing claim that everything is an Xbox. Now they're rationing the actual feature that made the claim true in the first place. So put yourself in the room where that math gets discussed. You've got a cloud gaming fleet that's confirmed to lose money on your most-streamed game, inside a division that just told the world it's "not healthy," sitting next to an Azure business that's one of the most profitable things Microsoft owns. If you're the exec deciding where compute capacity goes next quarter, cloud gaming servers that lose money per session are not where you fight to keep headcount and hardware. They're the thing you quietly right-size while the AI/Azure side of the building asks for more racks. Fifteen hours a month isn't really a customer-experience decision. It's a company deciding, as quietly as it can, which side of the business gets to keep the infrastructure. Or, tl;dr version: cloud gaming is a costly scapegoat for a much bigger problem — years of bad management and warring internal priorities that had nothing to do with server bills in the first place. It's easier to point at "cloud costs grow as usage grows" than to say "we bought too many studios, funded too many mandates that didn't work, hiked prices when growth stalled, and now every underperforming line item is getting squeezed to make the org chart look healthier than the decisions that built it." Cloud gaming didn't do anything wrong. It's just the line item that was small enough, and unprofitable enough, to cut without anyone at the top having to answer for the bigger bets that actually put Xbox here. Where This Fits If you've been reading the Insert Coin thread for a while, this will feel familiar in shape even if the specifics are new: a feature gets rolled out broadly and marketed as a value-add, people build habits around it being unrestricted, and once the usage numbers are in, the unrestricted part quietly gets walked back and re-sold as a paid add-on. It rhymes with the free, ad-supported cloud tier Microsoft's been testing with Insiders since July — one hour per session, five hours a month, ads up front — which in hindsight reads less like a separate experiment and more like Microsoft quietly figuring out what hour caps the market would tolerate before rolling them out everywhere. It doesn't make this wrong on its face — cloud compute genuinely isn't free, and "we're capping the thing 96% of you don't use enough to notice" is a defensible way to plug a margin hole. But it's worth naming plainly for what it is: another case of a company selling you unlimited, letting you get used to unlimited, and then discovering — right when the balance sheet needed it to — that unlimited was never really on the table. Got thoughts? Find me on Mastodon at @ppb1701@ppb.social Part of the ongoing Big Tech's War on Users series, filed under Insert Coin.