This post is an excerpt from the recent ARtillry Intelligence Briefing: ARCore & ARkit: Accelerating Mobile AR.
Besides scale, what are ARCore and ARkit’s qualitative attributes? Their biggest value is perhaps democratizing advanced AR. Developers don’t have to build AR technology from scratch, nor rely on physical barriers like markers or depth cameras. They can focus instead on end-user experiences.
ARCore and ARkit employ similar computer vision technology for mapping (the “M” in SLAM), which scans an area on which to infuse graphics. That includes horizontal plane detection, localization, motion tracking and light metering for realistic shading. These work in tandem to achieve true AR.
These methods approach the AR capability found in the Tango platform (ARCore’s forbear), which required depth-sensing cameras. ARCore and ARkit instead use surface detection software with single-lens RGB cameras, allowing them to reach the market scale explored above.
But perhaps more important than their similarities, how do ARCore and ARkit differ? That question applies to technical strengths, as well as go-to-market strategies and positioning. On both counts, ARCore and ARkit each carry its parent’s DNA, and each advance its parent’s core business.
For example, Apple’s app-centric paradigm is reflected in ARkit’s delivery, while Google’s web-orientation will shape ARCore’s stated web delivery goals. Meanwhile, Google has a technical head start with the years of work it’s invested in Tango, but Apple has more control over the hardware.
That software/hardware integration has always been Apple’s strength. In this case it can directly govern the camera optics and sensor calibration that support ARkit apps. Google is hoping baseline smartphone standards evolve, but for now it’s relying on the high-end S8 and Pixel (its own phone).
The following sections detail these differentiating factors, touching on technical factors but focusing mostly on strategic market positioning.
On technical measures, ARkit’s advantages include the vertical integration of hardware and software mentioned above. It can build hardware specifically to align with its software goals, and vice versa. It can also distribute software updates faster and to a smaller, more manageable range of devices.
This results in a greater portion of the iOS device base that runs the latest software, which has lots of advantages in terms of new feature adoption like AR, and security. That compares to the Android world, fragmented into several phones that rarely run the latest OS version, as quantified earlier.
Apple could also have greater appeal to VR developers – a key success factor as explored later in this report. This could stem from Apple’s app-approach. Compared to Google’s web AR, apps can provide a more structured revenue model, such as app sales, advertising or in-app purchases.
ARkit also has a speed advantage in being first to market. Its June release predates ARCore’s launch by months. Though that may be diminished by the long life span of mobile AR, there will be a slight early-mover advantage in attracting developers that become invested in ARkit.
Apple also owns an integrated software stack. This can limit developers to specific tools (though Unity has announced ARkit support). That development stack – including Metal and Swift – can create more functional and elegant AR graphics and experiences that run on ARkit.
To be fair, Apple doesn’t force these tools on developers but advises their use. This could still slow down ARkit app releases through a learning curve in the developer community. So a meaningful volume of ARkit apps could be 8-12 months away, partly negating Apple’s early-mover advantage.
Though ARkit has a slight advantage in being first to market with developers’ interest and invested time, greater developer attraction will ultimately come from platform reach. There, ARCore has the long-term advantage. As quantified above, ARCore’s installed base will start slow but grow fast.
ARkit’s technical capabilities and requirements likewise have a flip side. Though superior in several ways, ARkit’s technical chops could be challenged by ARCore’s flexibility. Importing graphical assets to ARCore will be easier than ARkit, given a more open approach and arsenal of developer tools.
In addition to working with 3D modeling tools and game engines like Unity and Unreal, Google has internally built developer tools that will help populate its immersive computing product line (ARCore and Daydream). These include Tilt Brush and the low-friction 3D graphics engine Google Blocks.
Overall, the immersive computing assets Google has assembled — more extensive and tenured than Apple — will support ARCore. Its visual search tool Google Lens will assist in object recognition. More importantly, Tango’s IP and human talent will be advantages realized and integrated over time.
Speaking of Google assets, one of its biggest advantages goes back farther than the last few years of VR and AR development. It’s the web itself, where Google has rooted itself for 15 years as the world’s search engine. That includes building a knowledge graph and search index.
This could play into an AR strategy by creating capabilities that are grounded in — and delivered by — the web. In fact, part of ARCore’s unveiling included discussion of WebAR, under which users don’t have to download apps, but rather visit mobile websites to summon AR experiences.
WebAR departs from Apple’s content architecture that’s rooted in apps. There is some evidence that an app-centric approach to AR could be disadvantaged, due to user friction. We’ve hit peak app-fatigue, given that most consumers’ monthly app install rate is zero, according to ComScore .
Beyond friction for users, apps are disadvantaged by their lack of interoperability compared to the link-structured web. This has always been the case in the app era, leading to movements like deep linking, but it could really handicap AR functionality by forcing it into non-linked silos.
“You effectively have all the same problems that a mobile app has,” Presence Capital partner Amitt Mahajan told ARtillry recently. “You have to convince someone to download it, and convince them to come back every day. All of the friction to get to that experience is still pretty high.”
Disclosure: ARtillry has no financial stake in the companies mentioned in this post, nor received payment for its production. Disclosure and ethics policy can be seen here.