Interesting, I gave Stadia it's first try as a result of this post. Tried the Worm Game. It confirms exactly my concerns about cloud gaming in general. The input lag.
The classic "worm" game is timing sensitive, gotta turn before you hit a wall or yourself, but also gotta turn as late as possible to ensure you don't leave empty space.
The input lag was noticeably bad, doing precise turns requires correctly guessing the input lag and clicking the turn command early.
Stadia isn't being shut down because it wasn't successful; it was no less successful than other services that are still up (like GeForce Now, or, for that matter, PS Now.)
Instead, it is my understanding that the point of Stadia was to give Google Cloud GPU instances a first-party "spot workload" to run when otherwise underutilized, since Google were building that instance type out ahead of confirmed business demand. They had all these idle GPUs just sitting around; may as well let people play games on them!
The moment the first GPU-backed generative AI models were open-sourced, though, there was suddenly tons of business demand for cloud GPU instances, to back all the "this[whatever]doesnotexist"-type services that were popping up. And "actually renting the GPU time to people with a business use-case for it" is much higher-margin than letting people play games with them. (That was the original intent of building out the instance class, after all.)
Stadia accomplished its purpose. The demand for the instances arrived, so Google didn't need Stadia any more.
This sounds like a retcon at best. Google wouldn’t have had to build a first party studio of the intent was to turn if off all along. I know the Eng VP that was there from the beginning (not to the end, tho!) and this is not his story at all. I think they honestly are trying to find new “surfaces” to give people access to compute.
Cunningham's Law at work. I've asked whether my hypothesis was anything like the truth numerous times now, but never received a knowing response. Phrase it one time as a bald statement of fact, though, and someone who can cite primary sources as to why I'm wrong will be summoned! :)
They recently refunded every purchase made on stadia when they announced it was going to close down, and they created 2 first party studios for it (along with deals with rockstar games etc.)
Did stadia itself make any money? Probably not. I'm sure the money they made from re-using those servers for their cloud/ai departments certainly paid of the dept though.
But derefr's explanation is the other way around. Google built the servers for the cloud/ai department and crated Stadia to reuse them. To me it seems that it couldn't have been profitable even if we say Stadia got the servers for free.
afaik stadia used custom built hardware from amd's instinct line. Almost all of their cloud offerings are nvidia based or utilize their own TPU's.
I can't say for certain that derefr's explanation is wrong, but from what it looks like they did create bespoke hardware for stadia instead of reusing their existing infrastructure.
I was skeptical on cloud gaming, but I also tried GeForce Now and had a great experience. No time spent downloading and installing, instant access to thousands of games, playable on older laptop without fans spinning up. I do think you need a solid internet connection for it to work well.
Unless the GeForce now servers were co-located to your local ISP, it's hard to believe. Even then, you're adding quantifiable jitter to the equation on the order of a few milliseconds. It's noticeable.
I played CP2077 few days after release date, no issues with input lag. On the other hand my ISP had trouble to keep stable 40Mb/s - so in peak ours it wasn't playable at all. It was a bit surprising at first but makes sense - even a second or two buffer for other streaming services is enough to ignore variations in connection speed - but gaming has no such luxury.
I’ve had the same experience. I could not tell the latency for GeForce now, and I’m using a municipal ISP.
I’m not really a casual gamer either. I’m still surprised that GeForce now doesn’t have an app for Linux. I switch to my Mac laptop to play stuff like Fortnite.
> I switch to my Mac laptop to play stuff like Fortnite.
Fortnite is about as casual as it gets. You won't rank very high in a twitch shooter like CS:GO (or keep consistent ranking) using a cloud gaming service.
I think the point being made was Fortnite tends to have a wider breadth of gamer than other games (casual to hard core). My partner has a group that plays, but it’s more about crazy fun For them. They’ve tried other games but found they just get dropped and lose too quickly to make it fun.
Most people playing any massive online game are casual gamers. That’s just math.
Just because one game has guns doesn’t make it less ‘casual’, which btw is a dismissive word to use in the context. Why are ‘twitch reactions’ less ‘casual’ than puzzle solving?
casual is the wrong way to gatekeep it. Someone could be a pro motorcycle racer with better reaction times than most professional gamers and still be a casual gamer. What you're really trying to test for is reaction times so just come right out and say that instead. Unless you're afraid of being beat by a filty casual.
I mean, in the case of multiplayer probably a lot comes out in the wash of optimistic updates and such, but it’s still very playable and I couldn’t personally notice the lag most of the time. Occasionally when my connection got bad it would have issues, but that was more noticeable because the resolution dropped significantly.
I think a problem with cloud gaming is that companies are being very secretive about where they have compute resources. Which I think is the wrong approach. they should partner with Comcast and say Atlanta is a Geforce Now hub. If someone is too far, they should just not let them access the service.
Most people put our ability to perceive network lag at 100-150ms. Human perceptual limits in vision is around 13ms, but that’s not exactly what we’re talking about here. 50ms is very very playable in networked games, I wouldn’t notice personally until 120ms.
I don't have a gaming PC or console atm so I can only use streaming (Geforce Now and Xbox Cloud) and I notice it but somehow you also get used to it like adjusting your gameplay to it. It's really not that bad (and I'm nowhere near any big data centers).
Anyways it wasn't the lag that killed Stadia but the monetization: buy the games full price + pay a monthly sub too. It's just plain bad. On Geforce Now you can use your existing libraries (Steam, Uplay, Epic) and don't even have to pay if you can stand the queues. On Xbox Cloud you have to pay a monthly sub (except for Fortnite) but they give the games too, pretty much a Netflix for gaming. Both has its merits, currently I dig Xbox Cloud more.
The free tier also came later and wasn’t well publicized: one of the lessons from Stadia’s failure is that Google giving it a bad reputation at launch squandered a lot of what they did later.
Another example was the refund at the end, which would have addressed one of the most common review points had it been policy. Here’s a representative launch review listing that under “The ugly”:
Same, it was by far one of the best performing games on the platform, and whatever stream post-processing magic was happening it looked stunning at 1440p. Likewise with RDR2. The latency was never an issue for me; granted I live near a major city and have a fiber connection.
Where "most people" is just the few people who actually used it? So basically, most people who like [unpopular thing] think that [reason that thing is unpopular] isn't a problem.
It didn't reach mainstream adoption because most people who care about games didn't think it was a good deal and input lag was a big part of the reason why. And people who don't care about games... didn't care about Stadia for obvious reasons. The target demo for Stadia was basically people who have a mild interest in gaming but not enough of an interest to already own hardware suitable for running games themselves. That was never going to be a profitable niche.
I've always felt the Game Streaming market was lower powered PCs, or something you can get "for the grandkids when they visit". People who have invested in having wired connections to their boxes (try pinging your wifi router sometime) and making sure their LAN and internet are quick probably also have invested in their actual systems. The cost of the games they play isn't the major driving force of their game choice.
I will tell you I 100% refused to buy anything from stadia because I was sure Google would cancel it in a year or two. You can attribute sentiment wherever you want, but Google failed because it was Google, not because it was a bad idea.
I used some nvidia thing that was out years before Stadia and it was amazing. Yes, I’d prefer crisper graphics and lower latency, but it was ridiculous for demoing games or doing casual LAN parties and such.
A lot of “supplemental” use cases for this technology.
I don't think it's the throughput, but the variance. There are so many places for something to go wrong between two computers, so doing something that needs a very tight latency budget is hard. Harder still when you don't have control over any of the parts connecting things. In a DC, everything can be built for qos and latency, but on a desktop you're using a porbably garbage router with a home OS running a million things.
I actually found the lag and compression artifacts between my Mac and gaming PC (which are connected by cable to the same switch) using the Steam remote play thing to be actually worse than Stadia.
a lot of product sun-setting from Google and other corp in this tech winter. Stadia is definitely handled in the best way one could hope for, IMHO it should be a text book example:
- refund the purchase
- listen to the user community, and act on it to avoid creating e-waste
- respect the employees and give them an opportunity to express their love of their work openly and officially
> Google deserves credit for how they have been handling the Stadia shut down IMHO
If Google had been as good at marketing Stadia as managing its shutdown, they never would have had to manage its shutdown.
Honestly, I think if they had been really smart, they would’ve made the game streaming a Youtube feature (including rolling the subscription benefits into YT Red. It would have been consistent with the way they’ve converged what used to be Google Play Music, and the way they’ve soft-converged Google Play Music and TV with Youtube.) It’s
In my opinion it's not a marketing failure. The combination of paying full price for games that require the service combined with Google's reputation for killing services left it pretty stillborn.
I think Google may have realised this which is why they're carefully managing the shutdown. It was too late for Stadia, but they don't want consumers feeling the next service is risky too by showing an intent to make it right when the shutdown happens
I disagree. Everyone that I know who has even the slightest interest in video games has heard about Stadia. They can't market their way out of the public perception around Google's long term commitments, nor can they change the laws of physics to eliminate input lag, nor can their marketing change the perception that paying full sticker price for games only accessible through their walled garden streaming service was a terrible deal.
They likely would have gained more popularity had they promised to refund all purchases in case the project shut down in less than a decade - in a legally binding way. This would have given customers a lot of confidence to invest time and money into their platform.
Additionally, they could have pulled in a portion of the existing PC gaming market if they allowed users to bring their own games. Having a service that would allow you to play your existing library while you're traveling or over at a friend's house would be enticing to many.
I wouldn't call this a marketing failure, however.
It's not a perceived vulnerability, it was a real, fundamentally unavoidable vulnerability, which is why we're here talking about the aftermath. This is why I disagree that marketing could have solved it.
So much. It's almost unbelievable. I never thought they eould have dedicated the resources to getting Bluetooth working (and possibly dealing with licensing/testing, since Bluetooth was never enabled in the first place—just BKE for initial setup)
It makes me wonder if the higherups are using this to counteract the worries that google will shut things down. I'd be less hesitant to invest in their services for fear of them shutting down if I thought I'd get my money back if they did.
I suspect they’re trying to avoid a class-action lawsuit. sunsetting software is one thing: it’s not so tangible and software licenses are legally slippery. but when hardware’s involved, you’re on much shakier ground
this idea that its to encourage future investment is predicated on Google thinking that the masses are paying attention to their refund policies. yes this kind of thing may be a big deal for people on HN or certain subreddits, but I doubt it’s a large % of the target market
There are lots of examples of support for hardware being dropped, the IoT space is full of it, likewise with non IoT devices, and I can't think of any examples of a company getting heat for it.
And google is very aware that their reputation for killing products hurts adoption of new launches.
perhaps I am underestimating how much Google takes that reputation into account, but I do still doubt there are many examples of a company with as many bags of cash as Google deprecating $50+ hardware that can only be used with its software, without offering a replacement product or service
The Stadia closure has left a controlled leak of fascinating internal Google projects. First it was cdc-file-transfer, "born from the ashes of Stadia," now it's the internal Worm Game used for testing that "the Stadia team spent a LOT of time playing." Love to see it.
Hopefully more to come. They ran all the games on Linux right? Linux gaming could take a step forward with the experience generated at Google and other studios that were porting the games.
This is exactly how Google needed to handle this situation, and they need to be praised for it! The shutdown of Stadia was completely predictable, but I am overwhelmingly surprised at them providing refunds to users for games & hardware and going the extra mile to ensure the hardware is not immediate waste.
As the cryptkeeper over at Killed by Google, I'm very happy with the outcome for Stadia users (myself included). However, I don't think this will move the needle of Google's 'killer' reputation as much as people think given all the skepticism about how long this product would last in the first place. The "death pool," and the people who voted basically predicted the timeline within a few days[1].
The irony is, if Google had made their plan for how to handle end-of-life for Stadia clear from the beginning, they might have got a lot more users.
I'm not sure if their strategy ever changed along the way, but when Stadia was first announced, it was a purely cloud gaming platform where you still needed to buy a license for every game. Not knowing what would happen to my purchased games was the main thing keeping me from trying Stadia - I'd have been happy to use it if it was a Netflix / Xbox Game Pass style service where you pay a flat monthly rate for access to a library, or if it was clear that any games you bought would be refunded if the service suddenly shut down.
As it was though, it seemed as though you could drop a ton of money on Stadia games, and Google might pull the plug the very next day, and you lose everything you just purchased. That worry was the deciding factor for me to never give Stadia a go.
Stadia required custom development work from game developers, my guess is that getting them on board was at least as challenging as getting consumers. It's a risk to tell developers "we're already thinking of how to kill this platform that you're investing in".
It's problematic for a company if they have to announce their EOL plans on product announcement to get interest from customers. Might as well not exist as a company at that point
It's problematic for a company to have EOL plans not because people will assume they will go out of business, but because they might not be able to actually adhere to those EOL plans resulting in a much larger mess than what would have happened if they stayed silent.
It's bad, but I'm not sure it's that bad. In B2B land, loads of contracts have a code-in-escrow clause just in case key people die or the company goes under. One of the key attractions of Open Source is that if the people building it die or don't want to support the tool anymore, you can keep going yourself.
call me cynical but I strongly suspect that rather than behaving honourably, or even thinking of future customer behaviour, Google are trying to evade a class-action lawsuit
the reason that this time they’ve seemingly turned a new leaf is that this time hardware is involved. once something is in someone’s hands and they actually own it, rather than a legally slippery software license or user agreement, you’re looking at something that can be very easily explained to a judge or jury
Strong disagree. For many weeks users have been thinking their controllers are garbage. They might still be. Let's see how well it works. Even if there's compensation I don't like to see stuff that's defective by design. https://www.defectivebydesign.org/
Perfect timing. I’m moving and just fished my stadia controllers out of the donate box after reading this - another day and they would have been gone. They’re decent controllers, so it’s nice they haven’t become yet more e-waste.
Next question is how good the Bluetooth chip is. The range of speeds even for the latest chips is vast. Since the controllers only used Bluetooth for setup, Google could have opted for a less powerful chip to lower the BOM and they might end up not being particularly useful for games over BT.
> I actually got Cyberpunk 2077 specifically on Stadia so I could play the meta-game of "Can I win Cyberpunk before the unaccountable mega-corpo revokes my access to Cyberpunk?"
They weren't "locked down" really, they just weren't bluetooth controllers; they connected over wifi directly to Stadia. This isn't removing shackles, it's adding new functionality.
Okay, I think i grok that. not so much locked down as they never built the capability. Although they did use BT for initial pairing with some devices so a certain amount of functionality was there.
From that starting point, allowing it to be a normals BT controller should require a minimum of additional work, but I think they probably had a reasonable business case to not do that:
We know they tried to implement to minimize additional latency, so I’m guessing they really did not want people using these as a general BT controller in case the user inadvertently connected to Stadia in that mode, which would have increased latency and reduced user experience.
Being able to plug in via usb for general controller usage was fine because increased latency there was not going to be an issues.
The Stadia controller is/was my favorite controller of any recent consoles (with my next-fave being maybe a PS1 variant). I've been using it wired with looong cords on other platforms because it just feels so good in hand. Very, very happy to hear this news. It'll be nice to go back to wireless.
Something I never understood - Why didn't Stadia push slow games harder? Games like Civ VI or Simulator games would've more or less hidden the input lag. Instead they seemed to be pushing fast-twitch games in their marketing.
I would guess the key factor for which games were pushed in marketing was the popularity & name recognition of the games, not any particular technical aspect. I think Stadia was also good enough tech to support the games they were marketing, at least for a casual player. I played Cyberpunk on Stadia and it was plenty smooth enough for me, but then I'm only a casual player.
Also my impression was that Stadia was aiming for "a console without the console" experience, and IMHO Civ and Sim games tend to be more PC focused (complex input requirements etc).
Nice of them to do that at least. It's a shame they're not open sourcing it for third party firmware though, if it can manage streaming games it must have a decentish processor.
That’s great news. Finally a way to use the controller from the free promotional Stadia kit (which also can with an already-useful Chromecast Ultra) that they sent me. Bizarrely, they sent it to me for free without any way of actually trying the Stadia service short of creating a fake Gmail account, because I had already apparently used my free trial and they wouldn’t re-enable a free trial despite sending me a $100+ hardware kit for free.
The trouble with cloud gaming is latency. Most solutions now just do a glorified remote desktop. It would require some rearchitecting, but why can't a client have an ultra-responsive frontend while farming out the heavy work to a cloud server somewhere? We're reaching a point where GPU's are becoming more and more expensive, so is a more distributed architecture the next step?
Weird, why were they even locked in the first place? I have a PS4 controller I connect to my computer; I never had to unlock it. Was Google afraid people would use their controller to... play games run on their own computers?
As others mentioned, but to add a bit more clearer detail, the controllers did not connect to the device you were playing on i.e your computer/tv etc, they used wifi to directly connect to the stadia servers thus avoiding an extra step and hence making them useless once the stadia servers closed down wirelessly, because they never worked like a normal controller (still would work wired though)
The controllers did however have a bluetooth chip that was used for initial setup, so you could connect it to the stadia app and put in the wifi settings. This firmware update google is releasing so that it will work as a normal bluetooth controller using this bluetooth chip.
There was no host console/PC, Stadia is a gaming system where the “host device” is in the cloud; there’s a local playback device (which might be a computer, or just a Chromecast stick) but that’s jist getting playing a audiovideo stream, not hosting the game. Goven the prevalence of WiFi for the playback devices, using a Bluetooth connection to the playback device to route controller input would add latency.
The architecture makes sense if you understand how the cloud gaming system is set up though. There's no general protocol for on board contoller wifi/SOC->cloud gaming services (yet). It's inherently a proprietary design.
It also didn't need a bluetooth chip but they added one anyway. It wasn't sold as a general controller so I don't really see the big deal. This is a nice good will gesture.
That's like creating clothes to only be worn a few times because you're targeting people who buy lots of clothes. A controller is an input device similar to a mouse and keyboard and ought to support standards so it can work with different devices. This "product perspective" seems to be that of IKEA and McDonald's.
You are assuming there was room on the device to support both at the same time. The amount of RAM/storage to support a Bluetooth stack may seem small but it's not for tiny devices. I'm sure they devoted resources to the WiFi path that the product was designed for instead.
From the other comments, it always had fully unlocked wired-USB support. And they did build Bluetooth support into the controller, and released software to enable that Bluetooth as Stadia was going under.
Still seems really weird not to enable it from the beginning--what would they lose?--but the made the right decisions during design and at the very end, at least.
They weren't "locked". You could play any game with them. They just didn't support Bluetooth at all. So you had to use them wired, or wireless over WiFi (for better latency when connecting to Stadia), but there is no standard for controllers over WiFi, so that part only worked with Stadia.
I can't blame them for not wanting to implement Bluetooth when it wasn't a requirement for Stadia. Bluetooth seems like a nightmare to support. But now they can release this as an unsupported best-effort update, and that's way easier. It probably won't work with some devices since they aren't going to test it extensively, but hey, it's free!
Another reason is that the Chromecast Ultra that shipped with Stadia didn't support Bluetooth. This was before the Chromecast with Google TV. So they had to make it WiFi to work on the Chromecasts at the time.
More than likely it they used some kind of proprietary communication stack for latency reasons or some such.
Other reasons may include:
- not having to deal with supporting other controllers on stadia
- not having to support their controller on other platforms
Not that it's defunct, they don't really have to deal with support aspect of things, and this is just a good will gesture so the controllers aren't paperweights.
Knowing that latency was such a problem that they had to take extreme measures like this to save even the tiniest bit of latency really should have been the end of this project.
Also seems like a benefit for being able to use "any" screen to game on. The screen doesn't need to know how Stadia works or interface with the game controller, it just needs to receive video and display it.
Stadia had a fundamental physics problem, not an engineering problem. Their "solution" is little more than a bandaid that highlights the fact that the problem exists.
Between USB polling rates, Bluetooth latency and OS processing, it's no wonder they wanted to avoid that entire stack and connected straight to Stadia. It's not a band-aid, those things together can easily result in 20ms of latency, which is an order of magnitude more than the latency of a fiber connection to Stadia servers.
What fundamental physics problem? Plenty of gamers, even competitive FPS players, found Stadia quite a good experience from a technical standpoint. Best in class, as far as streaming game services go.
Or you know, they could just avoid a whole class of compatibility issues due to not having to connect to a local device (which could be any of the mobile phones, Chromecasts, Android tvs or PCs with any of the four major OSes).
What's extreme is having to connect a peripheral directly to the internet because you need to eke out every last millisecond of latency. The fact that they felt this was necessary is directly related to the reason it failed; because the whole system was inherently susceptible to latency unless you lived very near to one of their data-centers. The product failed mostly for this reason. Everybody responding to me is acting like it was an obvious solution to the problem, but it wasn't a solution to the problem at all. It was an absurd bandaid placed over a gaping wound.
I don’t know why this Bluetooth thing triggers you so much. Only you seem to think that Google spent all of their engineering and marketing towards that one feature.
It added few ms latency to a product that requires low latency. It was easy to avoid it (having controllers talk Wi-Fi on top of Bluetooth is hardly a rocket science). It was likely one of many small projects to improve it.
I don't think it was your other reasons. You could use other controllers on stadia- they even made it a selling point that the only thing you had to buy was the game. And when connected via USB stadia controllers worked fine with Steam games (at least on Linux).
Another possible reason is that it wasn't seen as a high enough priority compared to other things they wanted to do.
In general, it would be because they can skip "things" that are required by a standard to support a wide variety of use cases, yet unnecessary for their specific use case. Skipped steps plausibly leads to higher performance.
Two things: They were never locked the same way your PS4 wasn't you could connect it wired and use it just fine as a regular controller. The second thing is that the stadia controllers used wifi as their source of connection over bluetooth, for various reasons Google said it allowed better latency with how they set up stadia.
So they're adding the ability to connect from bluetooth with this. In the same way before Sony added driver support to Dualshock controllers on PC, you couldn't use them wirelessly. They're not really "locked" more than the bluetooth support just wasn't there.
> So they're adding the ability to connect from bluetooth with this. In the same way before Sony added driver support to Dualshock controllers on PC, you couldn't use them wirelessly.
I only use it on Linux, and I suspect Sony didn't contribute the driver I use. Actually, I don't think PS4 controllers work wired at all, I believe they always use bluetooth and the wire is only for charging. AFAIK anyway.
Very odd that they decided to use wifi instead of bluetooth initially. Between this and the time they spent writing yet another file transfer tool, I'm getting the impression that Stadia was a very unfocused project, trying to do things differently (resume driven development?) when they should have been using off-the-shelf solutions and standard approaches for everything possible.
There are alternative drivers out there (for use with knockoff controllers that try to replicate the offical device IDs, for example), but if you haven't done anything special you're probably using the Sony driver.
I have a PS4 controller, not a PS5 controller. But I guess they did this for the PS4 controller as well? I would not have expected that from Sony of all companies.
Edit: As far as I can tell, the PS5 driver is hid-playstation while the PS4 driver hid-sony. hid-sony has had commits from at least one Sony employee since 2017, but doesn't seem to have originated from Sony.
The controller communicated directly with Stadia, avoiding the latency of having to push it through your machine's input stack and only then sending it to the servers.
The kit was being subsidised. At several points, they were essentially giving away starter kits of controllers and chromecast ultras.
I guess they figured it was concerning enough that people were going to pick them up just for a free Chromecast Ultra without also giving them a free controller for Steam too.
Fun fact: I ordered a Stadia "founder's edition" at launch for actual money, an order which Google later cancelled unilaterally. A couple weeks/months later, they emailed me a promo offering a similar kit for free (just pay shipping).
I've no idea why they didn't make money and had to shut down.
> Weird, why were they even locked in the first place?
“Locked” is a misnomer [0] ; they didn’t work as bluetooth controllers at all; because it waa a cloud gaming system, bluetooth to a local device that then had a connection (usually WiFi) to the clous backend would add latency, they used WiFi to connect directly to the cloud server, and bluetooth to coordinate with the device that would get the AV stream of the game.
> Was Google afraid people would use their controller to… play games run on their own computers?
They supported that via USB, so they couldn’t be too afraid of it.
[0] EDIT: or, at least, can give the misleading impression that BT controller functionality was active but locked to a specific use case.
...they weren't already? Is there a long German word that means "when a company lowers your opinion of them by announcing something you thought they were already doing"?
Google is the Netflix of something. Like Netflix is now known for cancelling series prematurely, for me Stadia was "that sounds cool, I should check that out" and then it was cancelled by the time I was seriously considering it.
More than likely, and knowing Google.. having it connect via WiFi meant that the controller can probably collect some form of telemetry on its users. Now that it's being shuttered, they can "remove" ('unlock') this feature.
https://news.ycombinator.com/item?id=34375251