Stadia is something that I think we all knew had a sort of expiration date on it from birth.
For over a decade, I’ve seen services that purport to offer streaming gameplay on cloud systems with a great user experience. There’s been a single problem that has banished them all to oblivion – if I can afford a fast and low-latency internet connection to use such a service, then I likely also have a system at home capable of playing the kind of games that Stadia offered, and if I can’t afford the type of internet connection that makes such a service work well, then I am a target customer remaining unserved.
Stadia has plenty of these problems, and I want to talk about them today, as Google has announced the disbanding of the Stadia Games and Entertainment division, which was going to use industry luminaries like Jade Raymond to deliver new, original titles to the service. While Stadia remains online and running today for now, it seems not long for this world, destined for the Google Graveyard of legend.
So what went wrong for Stadia?
Well, first and foremost was that it had an identity crisis. Streaming services can make a business out of offering players a value proposition (like a base subscription fee that includes game access) or by offering players with iffy home PCs a chance to play a standard purchased game in the cloud. Playstation Now is a good example of the former, where a subscription fee unlocks the full library of games along with any paid multiplayer modes that are locked behind PSN on the console, and GeForce Now is an example of the latter – you buy the games with a standard Windows PC license that you own, with the free tier offering limited play sessions and reduced machine quality, with a paid plan offering extended sessions and an upgrade to the hardware in use, including the ability to leverage RTX features.
Stadia was the worst possible compromise between these – offering only a Pro-subscription that offers free games from a limited selection, and otherwise requires that you buy your games at full price through Stadia, with a Stadia-specific license that isn’t transferable to another platform and requires the developer to specifically build and publish a Stadia version of the game you want. Sure, there is a free Stadia that functions sort of similarly to GeForce Now, but the difference is clear – GeForce Now leans on existing games from existing marketplaces and if the service gets shut off, you can still play them, while Stadia makes you buy Stadia games, build a Stadia friends list, and when (at this point, come on) Stadia is shut off, there does not seem to be any plan to ensure you can play those games, making the money spent on Stadia entirely wasted.
The one thing Stadia offered that I thought would be cool and even could be a good future state for gaming was the way in which it enabled instant-on streaming modes for social gameplay. There was a demo shown at a Stadia event that had players playing together in completely different roles in the same world, working to a common goal and able to communicate. One player was a flying POV, able to scout ahead and see what might be coming for the other player, who was landlocked. There was also the pretty-cool concept of state sharing, being able to pass your game off to a friend to show you how they would play through a certain difficult segment and then allowing you to play it, both of you watching in turn as the other played. There were also some streamer features that would have been great additions to YouTube streaming of games, like allowing viewers to join into games played via Stadia.
However, none of these features ever really came to be. State Share only exists in beta as a part of Crayta, a game that allows you to…make a game. None of the titles designed as proof of concept for these features made it out, and with the death of SG&E at Google, none of them are likely to resurface at any point, at least not in the Google-envisioned forms.
Instead, what could have been a unique platform with special and different games wound up being a dumping ground, where Google spent millions of dollars bringing standard AAA developers around with their standard fare. You can use Ubisoft’s $15 a month game service on Stadia, unlocking a library of titles for play with a monthly fee to Ubi (and you’d need a separate $10 a month to Google to get them up to proper 4k HDR video quality and surround sound), and the free titles that are on Stadia are just existing titles like PlayerUnknown’s Battlegrounds, Hitman, and various others. Stadia did have one of the smoother launch versions of Cyberpunk 2077, ironically, but that’s really about it. Instead of being the game service of tomorrow with unique and fresh gameplay concepts, Stadia is just another platform you can use to play Assassin’s Creed 10 or Military FPS 2,683 (this joke was working for me until I counted and Valhalla is the 11th mainline AC game, not counting spinoffs and mobile titles, and now I’m just baffled). I don’t even say that as shade (well, maybe for the constant stream of pointless military shooters that often fail to distinguish themselves from each other) but because everything that I actually wanted out of Stadia as a service was never delivered, and probably never will be at this point.
Further compounding matters was the confusing nature of the launch. Google’s standard leadership envisioned a Gmail-styled rollout, building hype through a limited audience, squashing bugs and growing the base as they went, but Phil Harrison, former Xbox chief, wanted a console-style launch where they sold some sort of hardware that allowed for preorders and to build hype that way. Instead, they landed on a weird hybrid of both, with the awful Stadia controller being a pre-order item that guaranteed early access to Stadia while holding off the full service launch until months past the controller became available. This, instead, became a messy point of confusion, as some thought you needed the controller, while others wanted to use Stadia with their existing hardware including controllers but didn’t have the early access, and thus the whole thing was a flop. Even worse, many people did not receive their early access codes with their controllers, causing the Stadia team to have to fix that later and leading to a wave of early adopters, people who were innately excited enough about the service to pay up in advance for a shitty controller, to lose their hype and discourage others from using the service.
With the dissolution of the SG&E division, we now have people coming out to discuss how things went wrong. Recent pieces in Bloomberg from Jason Schreier and a piece from Wired via Cecilia D’Anastasio show that a lot of employees already knew the writing was on the wall. The launch state of Stadia was charitably described as a beta, missing key features that the early announcements had hyped and presented as cornerstones of what made Stadia different. The controller, supposedly able to be lower-latency by connecting to Wi-fi on its own instead of running through an intermediary device, did not offer a meaningful latency reduction, with most testing showing it on-par with hardware connected to the host device playing Stadia games. Either way, Stadia’s input lag was abysmal, with average metrics in the 80-100ms range, meaning that at 60 FPS locked, a game would be receiving inputs 4-6 frames behind, which is absolutely not workable for games like fast-action shooters and fighters. Jade Raymond had been quoted as needing 5 years to really create the kind of titles that would make a service like Stadia, but SG&E, formed in March 2019, didn’t even make it to two years before being shuttered.
But yet, there is a potential next step for Stadia besides a slow, painful death and joining the other products, good and bad, in the Google Graveyard.
Stadia As White Box Cloud Gaming
As gaming expands via the rise of mobile and the vast majority of gaming devices are setup with always-on online connectivity, there is an interesting possible fate for Stadia.
Nintendo actually rises as one of the more likely candidates for a service that Google could offer via Stadia. A very small but growing number of titles are being brought out unexpectedly on Switch – games like Assassin’s Creed Odyssey (in Japan at least), Control, Resident Evil 7, and Hitman 3 are being sold on Switch as “Cloud Versions.” The game locally is simply a streaming client that connects to a cloud service of the publisher’s choosing, where the actual game visuals and audio are processed and sent to the Switch. This works really well for a ton of reasons on Switch – it grants publishers a larger audience for popular titles, gives Switch-only gamers access to titles they otherwise wouldn’t have, and reduces the amount of work needed to port such a title – the code of the actual game doesn’t need to run on the Switch itself and can just be a PC version with Switch button-mapping and slight adjustments to the game UI.
Stadia could be an easy white-box way for developers to offer more Cloud Versions of games on mobile platforms or the Switch. It already has a worldwide infrastructure, fast PCs designed to run games efficiently and with a degree of eye candy, and while the lack of the Stadia features promised for launch is a big deal to Stadia as a platform, if Google is just selling the infrastructure to developers to use in this way, no one has to even notice that lack of extra features. Hell, you could even allow developers access to those features in their own ways, if the demand is there. In this use case, Stadia being a dumping ground for AAA titles goes from being a weakness to a strength, since those are the titles most often incapable of running on the lesser Switch hardware!
Now, will Google go that route? I’m not sure. Google does offer things like G Suite, where their infrastructure is leveraged by other businesses to provide things like email, cloud office apps, and the like. The idea of extending Stadia to such a product is a sound way to pay off the investment made, which it should be noted was substantial. Google didn’t just buy off-the-shelf hardware, but instead bespoke Intel CPUs, AMD GPUs, with a shared memory infrastructure that used the HBM2 onboard the AMD graphics cards, and they bought a lot of this hardware to install in nearly every Google datacenter worldwide. Recouping the costs by offering it to developers for use in Switch or mobile “Cloud Edition” games would be smart. On the other hand, the AMD GPUs in use are strong compute products, and they could easily switch to using them for a variety of cloud compute projects and cut their loss on streaming video games.
Either way, Stadia feels like it is on a journey to an end. Whether it continues on without the branding is anyone’s guess, but I think we can all agree – it had a lot of potential that was unfortunately and predictably handicapped by launching a barely-competent, bog-standard streaming service first.