What the next few years look like for VR

“When is virtual reality going to take off — or fail?” Whether I’m talking to founders or other investors, most conversations I have regarding virtual or augmented reality eventually turn to this line of questioning.

Rather than making something up about where VR is on the hype cycle — which is descriptive, not prescriptive, so don’t assume you can use it as a guide for timing the market — I think it’s helpful to look at the specific hardware products that have been publicly announced and how well they might do — and where their relative successes will push the ecosystem.

Mobile VR

Two years ago, Samsung put a Gear VR device, a headset with a touchpad, lenses, and additional motion sensors that you slot a Galaxy phone into, on the market. However, Samsung treated it as something only for very early adopters until the second generation headset launched in late 2015. While we don’t get many updates on its sales performance, Oculus did announce that during at least one period in 2016, there were 1 million monthly active users. We assume Samsung has continued to sell more headsets since that update, but it’s difficult to get a sense for engagement over time. I don’t know if your typical Gear VR owner uses it once a week, once a month, or sticks it in a drawer after trying the top games and apps featured on the Oculus app store.

I have mixed feelings about the $99 Gear VR. After so much time with higher-end devices like the Oculus Rift or HTC Vive, I’ve lost interest in the more limited experiences a mobile headset delivers. It actually has very little to do with the graphics a phone can deliver; what’s missing is positional tracking, the jargon-y way of saying “you can get up and move around in a virtual space.” Gear VR, like Google’s Cardboard and upcoming Daydream platform, only tracks head-rotation today, effectively capping the sense of immersion users experience. But even as I consider this, there’s a little voice in my head arguing that thousands of people line up to try the Gear VR at game- and tech-industry events, and the limited capabilities might not outweigh the form factor’s advantages for casual usage at home or as a distraction on planes and trains. It fits in a backpack or large purse and takes seconds to dive into.

Speaking of Google’s efforts, Cardboard showed people that VR works on mobile phones, but the platform hasn’t led to the development of experiences that people would engage with every day. Between inconsistent experiences from the phones driving the apps, and the fact that some headsets required you to hold them to your face, Cardboard didn’t lend itself to becoming a sustainable developer platform, even if millions were shipped.

Daydream is a much more interesting effort. Google has effectively baked the kinds of optimizations Oculus did for Samsung’s phones — to squeeze a decent VR experience out of them—into Android itself. It also made a reference headset that Android OEMs can use to deliver their own Gear VR-style device to pair with their flagship phones (~$400 and up).

Google is also requiring phones meet a certain spec in order to be Daydream-ready. Unlike the Cardboard experience, Daydream apps will at least be delivered by a phone with real oomph and a “low-persistence” display to render scenes without nausea-inducing stuttering. And Google’s Daydream controller (kind of like a Wii remote with a touchpad) seems like a much nicer solution for mobile VR input than the touchpad on the side of the Gear VR.

The absolute best case scenario for mobile VR is that companies like LG and HTC feel compelled to bundle Daydream headsets with all of their flagships next year, exposing tens of millions of phone buyers to the diversity of games and experiences possible even on these more limited devices.

What’s more likely is decent sales numbers for most of the big OEMs (say a few hundred thousand each) and numbers comparable to the Gear VR for Samsung’s Daydream headset. Google’s upcoming Pixel phone serves as the highest quality but modestly selling device that puts the platform’s best foot forward, not unlike what Google’s Nexus devices have been for Android phones and tablets.

Intel, Qualcomm, Oculus, and startups like Occipital are all working on ways to bring positional tracking capabilities to mobile VR headsets—that is, both headsets powered by a smartphone and those made with smartphone components that don’t require a separate device to function. If those solutions work well sometime in late 2017 or into 2018, the options available to those who don’t want to commit to a PC or console will get significantly better.

Console VR

Moving up in price and capability, Sony’s PlayStation VR headset is what the majority of people who get to try what some enthusiasts consider “real” VR (again, that’s with the capability to move around a virtual environment by getting up and moving your body) will use in 2016–2017. From talking to other investors, entrepreneurs and journalists, I get the impression that Silicon Valley isn’t giving Sony enough credit for what PlayStation VR is going to do for the ecosystem overall. The hardware itself is somewhere between what Oculus offered to developers in 2014 and what they sell to consumers today, but with a real slate of games and media apps to take advantage of that hardware.

What’s really exciting though is the huge install base of 40–50 million PlayStation 4 owners (the hardware needed to power the PSVR headset), who might pick up a headset. PlayStation VR sold out at GameStop faster than any other hardware ever (though that’s likely because of very limited supply), and I wouldn’t be shocked if there’s a Wii-like shortage of PlayStation’s headset from launch into the spring. Over the next three or so years, I’d guess PSVR will sell somewhere between 10 and 20 million units.

Microsoft hasn’t announced a VR headset, but they’ve declared their intention to power VR experiences with a new Xbox coming out late next year. Codenamed “Project Scorpio,” the game console is supposedly going to have 6 teraflops of graphics compute, enough to drive the headsets available for gaming PCs today. That means it’s probably going to be powered by a new AMD APU (accelerated processing unit), basically an x86 CPU and decent graphics processing unit on one chip, based on AMD’s new Polaris architecture. If they could bring it to market for $499–599, they might be able to deliver the high-end experience of 2016 in a friendlier package that slots between the price of PSVR and the new high-end PC setups.

So which headset will we hook up to the new Xbox? Well, Minecraft is available on the Rift and Gear VR. You can use an Xbox controller with either Oculus headset. It’s possible it’ll be the Rift, maybe after a price cut. But no one likes sharing a platform, so there’s always the possibility Microsoft brings the HoloLens’s positional tracking to a VR headset.

PC VR

Connected to a PC, the Oculus Rift and HTC Vive will continue to push VR forward. On really expensive, powerful desktop computers, the visuals inside those bulky headsets will continue to improve, at first through better rendering and later with higher-resolution screens. If we’re lucky, Oculus or HTC and Valve will introduce wireless streaming from your desktop to the headset, eliminating the awkward cable that’s a bit too easy to get tangled up in. Eliminating the cable seems like a strong step toward VR becoming something we can just dip in and out of, like a book.

Looking at Steam hardware survey results, the bleeding-edge PC userbase that would buy ~$800 headset bundles—assuming Oculus and HTC continue to push specs up at a high price tier — is probably about 1–3 million gamers. That’s assuming there’s somewhere between 100 and 150 million monthly active users on Steam and that the bleeding edge is represented by those buying the newest, highest-end ($400–999) GPUs from Nvidia and AMD. So if PC-based VR is going to take off, it’ll be with those who don’t have the cutting edge chips, which — thanks to the aforementioned GPU makers—is becoming an increasingly large subset of PC owners.

If you don’t mind turning down some of the fancy visual settings, smaller PCs will be practical VR rigs. Today’s backpack VR PC becomes 2018’s bulgy pocket VR PC. It won’t be that long before even laptops as thin as ultrabooks today stream lightweight experiences to your headset without breaking a sweat. I still think that this segment will take longer to grow than PSVR and whatever happens with the Xbox, but that’s because those consoles have audiences who want immersive games, and VR headsets crank what gamers want (for some genres) up to 11.

So what does any of this tell you about who should or will be building things for VR? I’ve made a few assumptions based on what I’ve been seeing:

• Small game developers teams (1–5 people) can probably make a case for targeting any of the VR platforms (mobile, console, or PC) mentioned above. Even with install bases of hundreds of thousands to a few million units each, if you charge $10–49 for a game that stands out you’ll probably be able to eke out a high enough attach rate to make the effort worthwhile, as the folk at Survios proved with the awesome wave shooter “Raw Data.”
• Larger game developers willing to experiment with VR are probably going to aggregate to either Daydream/Gear VR or PlayStation VR over the next year, depending on how well those platforms do. The high-end PC space probably won’t be attractive to big studios without subsidy from Oculus or Valve themselves.
• One could argue two paths make sense for app developers — if you’re working on something that’s primarily social, target everything and try to build a strong network. If you’re working on a tool, it probably makes sense to target the high end to figure out the best possible interactions given the Rift and Vive’s capabilities and ride the curve of room-scale positional/hand tracking coming to consoles and mobile to build the network. A strategy of staying lean makes sense here, as it’s going to be a while before the install base numbers will make the math work without charging for things.

Why it could take longer

I’m a huge VR enthusiast, and so I generally find it easy to suspend disbelief about the space — I think virtual reality can provide extremely compelling experiences, and that millions of people will eventually agree.

But the vast majority of consumers, even gamers, haven’t tried virtual reality even once. If you’re a bit snobbish, you might even think that those who’ve tried the majority of headsets already out there haven’t seen “real” VR. If you haven’t tried it before, you might think it’s isolating, or too dorky to use in public like you use your phone.

On the first point, VR can be isolating in games and apps where it’s just you in there for extended periods of times. But I think VR will typically be a social experience, putting you in the same virtual spaces as friends, colleagues, and strangers at times when more typical telepresence or communication just wouldn’t do. On the second, I think VR will primarily be used from your home, the desk at the office, or while stuck on a bus or plane — not while sitting across from someone at a restaurant. But until people have tried VR even once, how are they going to get a sense for that?

Other factors could lead to virtual reality not “working” yet — the quality bar for typical experiences might be a few more years out, when there are 4K phone-sized screens in headsets or images are scanned directly onto our eyes. Or maybe VR experiences are compelling, but we just haven’t seen the form-factor people find appealing and won’t until Apple comes out with whatever their face computer ends up looking like.

Even if you’re a VR true-believer, it’s important to keep these factors in mind. But I tend to think that VR’s early foothold among gamers and developers will sustain the ecosystem’s momentum until we get the killer apps and devices that really nail look and feel and bring the category mainstream.