Last week, Google teamed up with the London startup Improbable to enable even the smallest developers to develop massive online games at low costs.

Improbable has created SpatialOS, software that enables small companies to create massive, cloud-based simulations for online games, mobile devices, and virtual reality projects. In doing so, it hopes to change the economics of connected games, and tapping Google’s own vast cloud platform will result in even more cost improvements and innovations.

SpatialOS creates a distributed computing environment that enables games to access the backend infrastructure needed to create incredibly complex systems, like the simulation of an entire city, said Herman Narula, the CEO of Improbable, in an interview with GamesBeat. SpatialOS is now available for game developers who create worlds that run on the Google Cloud Platform. With this new program, Google and Improbable will partner to subsidize access to Improbable’s SpatialOS platform. This partnership aims to encourage innovation by providing qualified game developers with subsidized access to Improbable’s world-building technology, powered by Google Cloud Platform.

With these moves, Google is effectively enabling developers to experiment with the technology for free.

Alongside this announcement, Improbable has launched its Game Developer Open Alpha for SpatialOS. Any game developer can gain access to the SpatialOS platform and development tools to test and experiment with it ahead of the full launch of the Games Innovation Program and the beta of SpatialOS in the first quarter of 2017. We talked with Narula, a computer science graduate from the University of Cambridge, about the deal and why it’s a significant opportunity for indie game developers as well as big companies.

Here’s an edited transcript of our conversation.

Bossa Studios' Worlds Adrift

Above: Bossa Studios’ Worlds Adrift

Image Credit: Bossa Studios

GamesBeat: I want understand why the Google Cloud partnership makes sense for you guys.

Herman Narula: The big reason for this is, fundamentally we’re trying to enable new kinds of applications and new kinds of games. That means that people need to take risks. They need to try things that may not work. By partnering with Google, that leaves us free to try things and experiment and explore, all the way through development up to the point of release. It’s going to mean that a lot more people than ever before can try these types of experiences.

The other thing we wanted to do is change the mentality around online games a bit. Historically those have been challenging to create and difficult to scale. By showing how you can play with the stack one day and be deploying things the next, doing all that really quickly and with the support of something like Google Cloud backing us — we think that’s an easy way for people to begin innovating, particularly studios who are normally quite confined in what they can build.

For the bigger studios, people looking to build really big franchises, this creates and enables a different mentality when they’re developing and testing their experiences internally. They don’t need to have a commercial engine to start building games. They don’t need to be cautious about how rapidly they want to build out. They can start scaling things straight away. It’s also about building an ecosystem. This is just the first of what we hope are more partnerships.

GamesBeat: Is Google subsidizing quite a bit here, removing a significant cost?

Narula: The way to think about it is, for developers building on SpatialOS, instead of us charging the developer, we’re making it free for developers who qualify for the program. Behind the scenes, Google is making free for us, which is what makes it possible. The subsidy is very large. I can’t go into exact details, but we’re going to announce more in Q1. It’ll make a real difference to developers. It completely changes the way they can operate, all the way to launch.

Worlds Adrift is using Improbable technology for a new MMO.

Above: Worlds Adrift is using Improbable technology for a new MMO.

Image Credit: Improbable

GamesBeat: So Google sees a payoff down the road. If you try something out on its system, eventually you’ll use a lot of its resources and that’s when you start paying?

Narula: Exactly. We’re also going to be very transparent about what pricing looks like. For most developers, it’ll be cheaper once they start building their own infrastructure. A lot of what SpatialOS enables is much more efficient utilization of resources. From Google’s perspective, this is a unique move, because they’re saying, “Even if you’re not directly working with us as a Google customer, you’re working on SpatialOS. We’re going to promote this innovation and subsidize it.” That partnership will evolve into other things as well. You can imagine some cool services being made possible through this partnership.

GamesBeat: I know a mobile or online game company can make a game and put it on top of Google Cloud now. They can offload a lot of infrastructure to Google. If you put SpatialOS in the middle, what does the developer get out of that?

Narula: The key thing is that you can now make new types of games that you couldn’t make before. The fundamental thing is that you can build an application that exceeds what a single game engine or server can do. We manage the dynamic semi-parallel distributed platform for this kind of computation that makes it possible to have a million things in the world, all interacting with each other in the same space, and have loads of users in the same world. Or have physics on the back-end that you couldn’t have before, or have high-frequency twitch gameplay. We’re very confined in our thinking about the types of games that are possible because of technology and because of creativity. We want to enable both.

Fundamentally it’s about being able to do more and being able to do it more quickly as well. If you do put up your own servers and try to build a new structure directly on Google’s compute engine, or any other provider, you’re really just renting hardware. You have to do all the work yourself — build up the infrastructure, manage it, build something that could take months or years to the point where you can create even a regular online game, let alone one that can do new things.

With SpatialOS, 15 minutes after you sign up you can begin sharing your game and developing and iterating, no matter how big it is. All the infrastructure is managed for you. You can use game engines you’re already comfortable with. You can plug in your own game engine. Game developers are amazingly innovative people. Why should they spend their time redoing the same things over and over again, dealing with the same limitations, when we can open that up for everyone? It’s a tradeoff in convenience and in new possibilities.

It’s about reuse and sharing, too. Developers can now share the code that they’re building in this ecosystem. There are already people sharing on the forums. It’s been three days, and we already have people getting together to give each other guidance.

GamesBeat: What about someone like AWS? Could it be a potential partner, too? Is there a particular reason to go with Google? Is it somehow optimizing its cloud for SpatialOS?

Narula: There are two reasons to partner with Google. The first is that they have a very common vision, very different from what a lot of players are doing. If you’re just making things easier, that’s not in itself a compelling proposition. But if you’re enabling new kinds of games, that is exciting. This first step is knowing that we embrace this vision and we want to do that. There could be other steps where additional services get put together, which makes it even more compelling.

Fundamentally, AWS and other cloud providers are providing hardware. It’s accurate to say SpatialOS abstracts that from developers. When you get your game running on SpatialOS you don’t think about where you’re running. But there are properties of Google Cloud and elements of our collaboration that make what we’re doing in building SpatialOS easier.

With all that said, it’s a non-exclusive partnership. Developers are free to interact or not interact with Google as they run SpatialOS. But the innovation subsidy runs on Google’s compute engine, all of which is abstracted from the developer and made very convenient for them to do. We’ll be announcing more details about this in Q1.

GamesBeat: With Google, I imagine that probably makes startups more comfortable working with you.

Narula: That’s been a good element, improving the general reception of everything we’re doing. But to be honest, it was more about encouraging people to do more things. I think other ecosystem partners will further improve that stability.

GamesBeat: As far as the imagination of developers, if I have a four-person team, and I’m thinking about making a pretty big game, traditionally I wouldn’t be able to even try to tackle any kind of large play space. How do you enable that, even with such a small team?

Narula: You’re outlining an important problem there. A few things are holding back small teams from building that kind of awesome game. The first is that normally, when you put a single-player game together, in a short time you can bash out some really rough gameplay to see what the experience could be like. You can find the fun, which is what developers want to do.

Doing that for an online game is almost impossible. You have to spend months building infrastructure before you can even start on gameplay that involves other people and allows you to find the fun on a scale that makes any kind of sense. We take that problem away. People are doing that right now. You can start finding the fun in an environment that has many people in it and has a lot of sophisticated stuff going on in the back-end to take care of things for you. You can begin prototyping more advanced ideas very quickly.

You also have to build out all the content. The kind of content you need to build often requires a larger team, which makes things more expensive. We can’t totally eliminate that problem. Creating art and building worlds will always take a lot of people. But we can make sure that if you do grow your team and do those things, you’ll be using those assets to do things that matter to the game. You can make art on top of proven gameplay.

We take over the financial risk of trying. You can bash out a game in month or two, like Lazarus, which is live right now. It’s a four-person team. They’ve had one programmer for most of its development cycle. It supports 3,000 users simultaneously in the same massive battle. They’re iterating and throwing in assets at the same time. They’ve made some modifications to the game to reduce the amount of content they have to build, but what they’re doing technically is very sophisticated and beyond what a small team could normally do.

We can’t take away the content problem, but we can allow people to innovate in other areas where it’s previously been very difficult. It’s a bit like what Unity did. We’re great admirers of Unity. They democratized one aspect of game development and allowed a generation of developers to innovate more quickly. We’re doing that for sophisticated online games. But more than that, we’re also enabling even the biggest devs to do cool new stuff as well. Some of the partnerships we’re going to announce are with larger, more traditional game developers.

GamesBeat: As far as all of the prep time, how have you guys spent your time to get to this point? It seems like you have some very good game companies trying it out, but only a few right now.

Narula: There’s much more going on than we’ve been able to announce yet. We’re hoping to increase the amount of public information and release things like dev diaries and examples of what people are doing. While we’ve been in development up until now, it’s been difficult for people to do things without a lot of support from us. That’s one of the reasons why we’ve been heads-down making this possible.

We also felt it was important to do something we could launch and release, something people could touch and feel and look at code with. When you think about things like new types of gameplay and crowd infrastructure, it’s difficult to just imagine what that could be like. We wanted to create something that let people see what that could be like. When we announce more studios and partnerships, you’ll see a lot more information about that very soon. You’ll see there’s quite a lot of traction already.

GamesBeat: It does seem like you’re in a space where it’s difficult work. Shinra Technologies shut down. MaxPlay couldn’t really get off the ground. What kind of lessons do you take into account there?

Narula: For us the play is a lot bigger than just gaming. We’re making an operating system that enables applications in all kinds of areas. The second thing for us is timing. A lot of what we’re doing wouldn’t have been possible even a few years ago. It’s stuff we’ve built from scratch or that takes advantage of new developments that have just happened very recently.

A lot of those services, while very ambitious and cool in what they were trying to do, I think suffered from two flaws. One, they tried to do rendering remotely, which makes it very difficult to build games that people can play, because the rendering is remote and then being streamed in. We don’t do that. We have a conventional game engine running locally, doing the rendering locally, and creating the same game experience for everyone without latency in interaction.

Also, they made it very difficult for developers to build in those environments. Building in a distributed system is very challenging for lots of people, even if they’ve been doing it for years. We’ve tried hard to integrate existing engines so that people can develop very quickly. There’s always the issue of cost, too. We’re hoping that the Google partnership and other things we’re going to announce soon as far as pricing make this very attractive to developers.

GamesBeat: Where do you think your customers will be? Are they going to be playing games on PC, on console, on phones?

Narula: What’s cool is we’re completely platform agnostic. We have VR games in development, like HelloVR. We have mobile games in development. We have PC games in development. Console is pretty straightforward. You can even mix gameplay experiences. If we’re successful in changing the way people look at these types of games, we can find our way into a situation where every platform is just one view into a persistent world you engage with.

You can put on your VR headset at home and wander through an amazing landscape, and then when you leave the house your phone is showing you updates on what happens in that environment. When you get to work you can pull up your browser and follow the interactions in a Twitch stream kind of way and interact with the world in a spectator mode. All of that is possible with Spatial today, and we have a diversity of developers working on different kinds of games.

I encourage you to check out Lazarus, because it’s so strange to us. We never imagined that someone could make a game like that. We didn’t create this technology to support a 2D shooter, but someone’s made a 2D shooter with SpatialOS that wouldn’t have been possible otherwise.

GamesBeat: As you guys get through the engineering part, have you changed things much, or do you think you’ve kept the same vision along the way?

Narula: If anything, the vision has become even bigger. We see how many different kinds of developers are willing to embrace it. Now we’re very keen to see how we can take this even further. We have some awesome new things in mind, like additional engine integration, and others that we’re working on. We want to broaden this out more and more. We’re also seeing how we can help ecosystem partners and carve out a greater space for everyone by being a point of integration. Like we’re doing with engines and services, we become a great way for developers to access those things and bring them into their world.

I wish we’d gotten here sooner. There’s probably more demand than we can conceivably deal with right now. We’re trying very hard to scale.

GamesBeat: Remind me how you guys are going to make money.

Narula: It’s usage-based. We’ve put this on our site and we’re going to announce more details soon. Basically, you’ll pay for usage. You won’t need to buy servers or pay anyone else. You’ll just pay for usage. We’ll be very clear about this with developers. You’ll be able to plan very easily what your game will look like. Our costs will go down over time for lots of reasons. Even now it’s cheaper to use us than it is to build your own infrastructure in many cases.

The cost is also tied to instantaneous usage of your world. Unlike other systems, where you’ll have to rent a bunch of servers and maybe no one will play, with SpatialOS you only pay for exactly what’s happening right now. You can get it down to a very low cost depending on your type of game. Lazarus is cost-effective and it’s made by a team of four people. Worlds Adrift, which is also made by an indie studio, can be cost-effective. That should give a good indication of what’s possible for even an ambitious project like that.

Worlds Adrift is in continuous alpha right now. It’s going to stay live, we think, or have very rapid alphas. They’ve announced that they’re going to be launching to early access in Q1 of next year. They’re completely on track. Lazarus is in open development and live as well, and they’re likely to be announcing alpha very soon.