In 2018 and 2019, I was working as a Product Manager for GeForce NOW, Nvidia's cloud gaming service. At the time, the service was in closed beta and still working on some challenges. How many blades should we buy? Where should we build new data centers? How do we get permission from publishers to stream their content? What should be the price point? Is the technology ready for mass-market? Who are the consumers of this product? I was working on the team that sought to answer these questions.
At the time, we also saw the rise of competitors in the cloud gaming space. We knew Amazon was working on what became Luna; Microsoft was openly demonstrating xCloud; and Google was starting to tease what became Google Stadia. Each company had its own competitive positioning: Amazon had Twitch, Microsoft had Xbox, Nvidia had GeForce, and Google had massive reach and Android. The GeForce NOW team took each competitor seriously, but it was clear to me that Google Stadia was doomed to fail.
I'd initially wrote this in January 2020, but chose not to publish it given my recent tenure at Nvidia. But in 2022, I can say that the three arguments against Google Stadia still seem true today:
- Developers won't port their games to Linux.
- Consumers who want to play AAA games already have the hardware.
- Stadia lacked compelling content offerings.
Developers won't port their games to Linux
Getting developers to port their games to Linux is very hard. It's a time-consuming effort that's been tried before (remember Steam Machines?). With so many different game engines, extensions, and development environments, there isn't an easy way to motivate an industry-wide investment in Linux gaming.
Why would Stadia require Linux? Because the Microsoft server licenses are expensive. While I won't quote numbers for GeForce NOW, these licenses are a significant portion of the cost to run a cloud gaming service. I suspect that, by moving to Linux, the Stadia team thought they could turn a profit by either paying developers to port their games, or doing the porting for them. They were wrong.
There is no profit for the developers to do it themselves. Each new supported platform introduces more and more costs to development. Every new game build has an additional platform that needs to be tested, every new patch has another store upload, and every microtransaction has another payments API it has to interface with. With Linux sporting only a 2.8% market share in desktop, it's just not worth it for developers to port their games voluntarily - they must be incentivized.
Consumers who want to play AAA games already have the hardware
If you are someone who's deeply interested in playing games such as Assassin's Creed or Red Dead Redemption, you likely already have the hardware. The consumers that engage with these types of games are more hobbyist than everyday consumer, and hobbyists like to spend money on their hobbies. How many people are there that want to play Darksiders 3, but just don't have the hardware to do it? I suspect very few.
Stadia also wasn't a hardware free solution, requiring you to purchase a controller and a Chromecast for $129.99 (at launch). Costing 2x the price of games at the time, this continued to stress the question: "Who is Stadia for?" For $70 more, consumers could purchase a used Xbox One or PlayStation 4 and accessed a huge library of free games.
Hobbyist gamers were also least likely to engage with Google's business model. The models offered by Nvidia and Microsoft - bring your own games - was and still is vastly superior. The BYOGames model lets gamers enjoy cloud saving, portable games, and their own hardware/peripherals (ie gamepad, headphones, etc). Consumers who think "I'll own a gaming PC or Xbox some day" are much more likely to commit to these established gaming ecosystems over an unknown one.
Stadia Lacked Compelling Content Offerings
The most surprising decision by the Google Stadia team was to launch without first-party content. Of its 38 launch titles, many were released in 2018 or earlier, and almost all of them had been deeply discounted at some point. None of the 38 launch titles were intense multiplayer games with deep gameplay and social engagement (like Fortnite, PUBG, League of Legends, Valorant, World of Warcraft, Rainbow Six Siege, CS:GO, etc.).
The "middle market" of compelling, single-player games is already stressed and in decline (from a revenue perspective), with subscription and live-service models taking off and offering massive value to gaming consumers. There's also been a huge surge in mobile gaming, with games like Minecraft, Call of Duty, League of Legends: Wild Rift, and others expanding their core gamer engagement onto the platforms "meant for cloud gaming."
Stadia's bet that people would still buy these games at full price was simply flawed, but their arrogance in assuming these games would attract consumers was absolutely stunning. When Stadia shut down their gaming studios in February 2021, the writing was on the wall: Stadia was giving up.
There is no doubt that Stadia was and still is an impressive technology. The controller itself was a brilliant idea to further reduce input delay by connecting it directly to the internet and not an access point, and was something I vividly remember giving the GeForce NOW team pause. The ability to play games in a web browser was also an incredible feat and motivated our own internal experiments at the time.
But no amount of brilliant technology can overcome a lack of product-market fit.