Google’s Augmented Reality Course Correction

Earlier this week Google made a huge strategic shift in its plans around augmented reality, announcing ARCore, a software developer kit that will enable Android-based AR apps in much the same way that Apple’s ARKit will enable AR apps in iOS. The announcement reflects a needed course correction by Google and has the potential to dramatically increase the number of consumers with access to phone-based AR around the world. It may well signify the company’s changing expectations around virtual reality, too.

Tango’s Tortured Dance
Before this announcement, Google seemed to be sticking with its pioneering but problematic Tango AR technology. Tango, which began life as Project Tango before Google moved to make it official, requires very specific, high-performance smartphone hardware components. The cost of these components and the power requirements of Tango made it very difficult for hardware vendors to build supporting smartphones, and at this writing, there are still only two shipping products. Google says it has been working on ARCore for some time, but assuming that’s true the company clearly wasn’t ready to show it at Google I/O in May. At the time, the company was still talking about Tango on stage, although the presentation was pretty lackluster, as I noted at the time.

The key difference between Tango and ARCore is that the latter will work without any special hardware. The core technology of ARCore focuses on three key areas: motion tracking, environmental tracking, and light estimation. The first uses the phone’s camera to understand where the smartphone is in the room and to measure changes as the phone moves within the space. Environmental understanding is the tracking of surfaces and objects within the space around the phone so that you can place AR objects into that space. And Light Estimation is how ARCore captures the light in an environment so that the AR objects have the right lighting, which is key to making them look more like real objects. I haven’t seen the technology in person, but the videos on Google’s site make it look pretty good.

Big Targets
Google says that ARCore will work on millions of existing smartphones, and it will roll out the technology first to the Google Pixel and Samsung S8 devices running 7.0 Nougat and above. The company says it is working with a wide range of Android smartphone vendors to enable it on their devices, and it expects to have ARCore running on hundreds of millions of phones by the end of the preview period. When Apple ships iOS 11 to its customers later this year, it will enable ARkit for 300-400 million existing iPhones and iPads. The impact of this one-two punch can’t be understated: AR is coming to the masses, and it is coming soon.

Ben Bajarin’s recent piece on this topic articulated the economics around Google and Apple’s ecosystems and the opportunity this represents for developers. What’s particularly interesting about this is that while Android obviously has a much larger installed base than Apple, ARCore will reach a significantly smaller percentage of that installed base then ARKit will within the iOS installed base. And, of course, iOS owners spend significantly more on apps than do their Android-owning counterparts.
Impact on VR?

Google has always been a company that has been willing to change direction, and even kill products when it’s clear the market isn’t responding to what’s on offer. Whether it happened before Apple announced ARKit or shortly after that, the fact that Google finally recognized that Tango wasn’t going to drive mainstream adoption of AR and moved to do something about it is commendable. The company says it learned a great deal from Tango that will make ARCore better, and I believe that. It will be interesting to see how the company leverages those learning in the market.
We should also pay close attention to how this shift impacts adjacent technologies at the company. At Google, I/O, the company’s virtual reality platform Daydream received significantly more stage time than Tango. But the company’s progress in this area, especially around Daydream-enabled smartphone launches, has been pretty slow. At its developer conference, Google said it expected to see tens of millions of Daydream-enabled smartphones in the market by the end of 2017, which is a number much smaller than its new target for ARCore. Google is a big company, and it can drive more than one smartphone initiative at once, but it’s hard not to look at ARCore and see the potential for the type of scale that Google really likes. It’s also worth noting that the Google file the blog post announcing ARCore under “Google VR” on its site.

However it all plays out, one thing is clear. Having both Apple and Google focused on Mobile AR is good for the industry, good for developers, and good for consumers. And while the early apps from both will focus on delivering augmented reality experiences to the phone screen, I continue to believe that eventually, the best experiences will appear on a set of glasses wirelessly tethered to the phone. With Google and Android now more fully in the mix, this outcome seems even more assured, and when it happens, it will do so at a much wider range of price points.