Instead of simply gluing a remote session to the internet, Stadia requires developers to change how they think about building games: they are now rendered in aggregate across servers, with game logic for things like multiplayer abstracted into a separate process, and rendered right at the edge in one of Google's existing cloud datacenters.

That means Stadia won't launch with every game in the known universe but likely only those optimized for the platform, so that it's able to provide a 'console-like' experience seamlessly.

Google has an interesting trick up its sleeve to reduce perceived latency as well: a dedicated Stadia controller that connects to WiFi rather than directly to your device. The controller tunnels to the datacenter instance you're connected to directly, on a separate 'pipe' from the video stream, further reducing the feeling that you're not really playing locally. 

Fully baked

All of this is wild, but what's so astounding in my mind about Stadia is that Google hasn't rushed this out the door: I was able to find sources that told me the company has been working on it since as far back as 2013, with efforts in earnest beginning in 2014, and many iterations being thrown out before that.

Google understood that launching Stadia (codenamed Yeti internally) half-baked would mean the gaming industry would simply snub it, so it waited, and worked on this for years until it was a leap ahead of any competition.

With game studios like Epic Games onboard and support already baked into the Unreal Engine, this isn't a technology demo or showreel, but a declaration that one way or another this technology is going to change gaming. Game streaming is likely to change a vast array of the industry's reliable variables: consumers buying consoles,  consumers owning games outright, how downloadable content works, how many players can play a game simultaneously, the business models of the studios, and so on. 

But, there's a sticking point in Stadia's mission to bring 'gaming to everyone' that went unsaid: internet access is inherently unequal globally, and streaming might shut out millions of players from the 'next' generation of gaming. about the same speed being able to achieve a 4K stream.

Google has many more datacenters than any other competitor, so it's able to get closer to more consumers than almost anyone else on the planet right now, but that doesn't necessarily solve the last mile where ISPs in many countries have a monopoly over cables and drag their feet over improving connectivity. 

Services like GeForce now require a stable connection of at least 20 Mbps to achieve a 1080p stream, and Stadia improves on that with about the same throughput able to achieve a 4K stream. The average connection speed in the US? A measly 18 Mbps. Stadia's debut ignored this reality, but acknowledged it in a subtle way: it'll be global at launch and notably available in Europe, where connection speeds tend to be much higher on average. 

Stadia's launch this week showed a service that is very nearly ready for the prime-time, but needs game developer support to shine: and that's why we got very few answers about when the public can use it (2019, supposedly), what games it'll ship with or what it'll cost. The demonstration was designed to woo developers, and lure them to Google's vision of gaming in the future, while getting the public excited about what that future might look like.

If anything, Stadia is a wake-up call for the gaming industry: here's a cloud platform that will enable you to build unique games at an unprecedented scale, and reach gamers wherever they are. What we're now waiting for is seeing how widely it's embraced.