Gaikai's David Perry
On 100m users, the Gaikai controller, Amazon and cloud gaming's 'dinosaur' moment
CEO of Gaikai David Perry is one of the most passionate advocates of cloud gaming in the industry. In this interview at Cloud Gaming USA, he spoke to GamesIndustry.biz about Gaikai's business plans, peripherals, bandwidth and modding concerns, and the potential for creating games that couldn't have been done before.
Q: One of the big concerns people have about cloud gaming that that a game could be retired out of the network and they won't have access to it anymore. What would you say to that?
David Perry: Well that's a great question. Digital means infinite shelf life. A great example of that is Kongregate's website. If you look in there they have like 260 pages of games, but you don't ever retire those games. They're always there if you want them. It's the same with cloud gaming. Cloud gaming just means in the data center there is usually a cache where the games are sitting and hard drive space isn't a problem. So I don't really see any need to retire games. If there's people wanting to play them at all we would leave them there. Today we have four terabyte hard drives. It's just not an issue anymore.
Q: Do you have any figures to show how many people would access to Gaikai?
David Perry: We technically just passed I believe around 10 million. We are targeting 100 million as quickly as we can possibly get there. We need to get above the reach of any single game entity in the industry a quickly as we can. If you were to put a game on a single website that's not going to have anything like the reach that we'll be able to have.
We're doing it three different ways. We're doing it through retailer websites -- we have a bunch more announcements coming out on retailer partners. We're doing affiliate websites. And we're changing the way in which we're going to handle affiliates but I don't really want to make an announcement on how we're going to do it yet as I don't want to be copied. But we have ways to do even more viral spreading of Gaikai.
When we first started Gaikai I remember saying to the publishers "what we really want to build here is a solution where you can have a meeting and say 'damn, some game just slipped a quarter so we need to have more people play this other game over here. Let's have a million people play that right now.'"
I've been starting to look into how we can make a controller at a sensible price as it's quite a complex device.
The ability to pull a lever and have a million people play your game is something that's crazy to even think about today. That's a very difficult problem, yet with cloud gaming that will be really quite straight forward. A normal way of thinking will be, "I want another million and another million."
It turns out that the number one way to get people to buy is trying your product, and it's amazing that game trials are being made so difficult.
Q: When do you hypothesise you'll reach your goal of 100 million?
David Perry: I'd say by the end of next year.
Q: One of the big benefits to cloud gaming is you can do it on multiple platforms. Do you think there will be a homogenisation of game consoles with hardware like Kinect that can't demo on a tablet or on a regular TV? Do you think it will be a problem that hardware developers like Microsoft and Nintendo are going in these divergent directions?
David Perry: Kinect is a funny one because it's actually made by a company called PrimeSense. So it's not actually technology made by Microsoft, so there's nothing to stop any company from putting a PrimeSense camera on your device. That said, all controllers: joypads, guitars, anything that could end up with a USB connector can plug into an awful lot of devices. Not all tablets. Not the iPad. But if you can get to a USB, that's good. If you can get to bluetooth, that's good.
So Gaikai has been working on a way to do what we consider to be incredibly seamless controller support, so we've done demonstrations where if you plug in a USB it actually tells the server that the USB's been plugged in and you'll see it reflected in the game in the cloud switch to joypad mode. So it's about as frictionless as it can possibly be. Meaning we've found a way to make it so even though this game isn't in your computer, it knows that a USB just got plugged in and it unlocks that mode in the game.
Q: Would Gaikai have any interest in creating its own peripherals then?
David Perry: We've certainly been having discussions on it. I think we will at some point make a Gaikai controller ourselves. Because I have my own ideas for what a controller needs to be. I've been collecting controllers to demonstrate my point on what needs to be done. I've been starting to look into how that can be made at a sensible price as it's quite a complex device, so we'll see. But for the minute our plan is to support as many things as can be plugged in or go through Bluetooth. Whatever your favourite controller is. If it's an Xbox controller, plug it in. It works on Gaikai today.
Q: You mentioned demos being the most prominent way for people to make purchasing decisions. So are you interested primarily in demos or are you more interested in selling full games?
David Perry: We are a company doing this for other people. We are not doing this for ourselves. So our goal is to serve them and deliver this kind of experience. If this was just Gaikai and Gaikai has down time that's our problem. But if I'm doing it for Electronic Arts or someone else and it's down, then it's my problem big-time and everyone's going to be calling. "What the hell? Your network's down!" So advertising to us seemed like the best way to get started. If your ads go down for a certain part of the day it's not as big a deal. It's our problem. Of course internally we'd be dealing with it. But if you just paid to play a game and you log in and it's not available for any reason you're going to get mad.
We worked out this strategy where we'd have two phases. The first would be all advertising and the second would be full games. We expected one to come after the other, but not quite this quickly so we're at a point now where the conversions that we're seeing on the partner sites are way higher than expected so now we want to go to full games. The problem with that is I need to offer Amazon-level "always up" service time.
We built a company like Verizon that doesn't take ownership of anything you're doing on the network.
That's the thing with cloud gaming. It starts off such a simple idea; you're getting data from the cloud. But its all this other stuff: running data centers, making sure everything works all the time... All these different services running together. So we're having humans testing it right now, but you can't do 99.5 per cent with humans. So we have to write all kinds of different test harnesses to make sure every single part of the system is tested using code and not with people. And these are the kinds of things we're dealing with right now. So from now through November is really just all a lot of engineering to do a lot of beta testing.
Q: So you don't have much say in what publishers do with it?
David Perry: If I built Gaikai as a company where we owned your customers, and your analytics, and a bunch of your revenue, I can be certain that within 18 months you'd call me up and say, "I just came out of a board meeting and we decided we needed to own our customers for digital going forward and we need all the data we own. We really don't want to be paying any more revenue to an external company, so we're pulling out." That would be what would happen if I went down that path.
We built a company like Verizon that doesn't take ownership of anything you're doing on the network. It's a bit like Amazon. I look at Amazon more as a competitor than anything else because they are supplying GPU capacity from one data center. I have 24. And so that's what we're doing. So as a company the money we generate is based on uses worldwide of cloud gaming.
If you think about it everyone is aligned. Our goal is for people to do it and have a great experience and come back for more. Our biggest epiphany is the quality of the service is more important than anything else and that's partially why we're going to switch from human testing to doing everything with software.
I can say this safely as a developer. As a developer you care so much about your product that it's got to just rock. The latency has to be as low as it can possibly be. The game has to run as fast as it possibly can, and it has to look as good as possible. And that is the service that we're building. It's expensive to do and we're investing very heavily in the infrastructure to make that possible and the net result will be very pro developer and at the end of the day those are the people who we have to make very happy.
Q: Would small indie developers still be able to afford your services?
David Perry: They can make up their own pricing. If I were an indie developer, I would swing for the fences. Every new platform comes out and there's a title that we see that we all want to experience. 3DO comes out and there's FIFA Soccer and it's the one title that makes you want to try that product. With PlayStation it was that dinosaur demo. When you saw that running you were like "oh my god, this thing's unbelievable!" So if you think about cloud gaming it needs that game. Something you wouldn't even think to bother with on a normal distribution model. It'd be something where you'd want to max out the memory. We're going to just bring this thing to its knees, and have that experience delivered to my little notebook.
It doesn't have to be an indie, but somebody's got to make that. If I were in the development business I'd say there's a huge opportunity to make something people are flocking to experience.
Q: The kind of game you're describing sounds expensive.
David Perry: Not necessarily. All the limitations you're always working around - building multi-level models or whatever tricks you need to make it perform in this environment - could really rely on the server being there for you. You could do much larger textures and push the hardware as far as it would go. At the moment the things we use are visual benchmarks like Heaven and things like that. They really show what the hardware is capable of. But its not a game that you can actually play.
Q: But would the cost for indies to put their game on your service be affordable?
If I were in the development business I'd say there's a huge opportunity to make something people are flocking to experience on the cloud.
David Perry: You pay as you go. There's no financial commitment. You just pay for the time you're using. Stage two is you tell me how many people you're planning to serve and you can low ball that and I'll cover the excess. You reserve that capacity on the network for a set amount of time. That is the lowest cost way of working with us.
Q: Would PC gamers still be able to mod games if they don't have them on their computers?
David Perry: Absolutely. Cloud modding allows you to add mods with a single click. If the publisher allows it that is. So modding in cloud gaming is a way better experience than having to do it manually by yourself.
Q: So if someone like Valve releases their source code, would there be a way to get it on your hard drive to tinker with it?
David Perry: There's a whole mod community out there and every now and again some mod almost becomes standard due to its popularity. That mod, if the publisher agrees, we can put onto the cloud and you can turn it on with a single click. They're not going to approve every one as some will have people running around naked or whatever. That's the way I'd expect it to work with cloud gaming. You'd try out the game and if the approved mods click you use them. There's no cheating or hacking.
Q: But what about to create the mod in the first place? Would there be an option to get the game locally?
David Perry: At this point it's just a developer question. Do they allow you to or don't they? We're certainly not going to stand in the way of that. It's a very, very passionate concern for a lot of people as gamers love their mods, so I think it's important.
I think a lot more people would try mods if it was just one click, then they realise "oh, this is better."
Q: Are you concerned about bandwidth restrictions that are becoming increasingly common in some parts of the world?
David Perry: As long as it's not crazy like terabytes, I think you're okay. In reality, the amount of bandwidth everyone uses goes up together, so what they're looking for are people who are using freakish amounts. If the world is doing more and more streaming and we rise together, which is what you see happening, then we're all using a lot more bandwidth than before and we're not all getting turned off. It's only the one per cent that stand out that the ISP's think are abusing the system that get shut off.
Q: What about cost to consumer for bandwidth? In some places people are charged by how much bandwidth they use and this could quickly become very expensive for them.
David Perry: If someone is in one of these plans where they have to pay for the actual data used then they would want to use the least amount of bandwidth they possibly can. We thought about that and said, "Why don't we offer a second option to do a cloud based delivery of the game?" So if they don't want to keep streaming for 100 hours and want it locally, what is the fastest, most convenient way to get that running on their system? We've definitely changed the rules on that. In fact, it uses less than downloading the whole game as we only download what they need.