Thoughts On XR To Date
A few thoughts on the XR/Spatial Computing sector (augmented reality, virtual reality, Xverses) to date:
(1) The major tech companies have spent tens of billions to develop XR.
My best estimate of the collective spend of Meta/Facebook, Apple, Microsoft, Amazon, and Alphabet/Google in XR from 2012–2022 is somewhere in the range of $100-$150 billion (USD). Meta/Facebook and Apple likely spent $35–50 billion each over this period of time (including Facebook’s acquisition and integration of Oculus).
Meta/Facebook likely has a current $10–15 billion/year burn rate in XR development for 2022. Apple’s XR spend should be similar this year, maybe even larger. This level of spend may continue for the remainder of the decade, as the major players race each other to dominate what may be the next major shift in computing.
To provide context, Magic Leap (the spatial computing/XR company I founded in 2011) has raised > $3 billion USD in over a decade. This is a significant amount of capital, but it is roughly what Magic Leap’s primary competitors (major tech companies like Apple, Facebook) may spend every 3 months.
(2) The XR/Spatial Computing market needs many billions to incubate, grow, learn, and develop.
Many categories in computing and consumer electronics have had a precedent or similar product already in mass use for decades. The typewriter and television combined to form the early PC (mainframe computers had been in use since the end of WW2). Headphones and portable AM/FM radios and casette players (ex: Sony Walkman) were already in mass use, so the movement to MP3 players like the iPod did not require a major leap of faith. Devices like Beats headphones and Airpods also flowed from decades of people wearing studio headphones and smaller wired or wireless headphones. The mobile phones of today can trace almost a century of innovation from the classic landline to the emergence of cellular/mobile phones in the 1970s. The electric cars of today are also well understood formats — we know what cars do, we know how to drive them, and there is a lot of sense in elminating a dependency on oil (for both environmental and national security/economic reasons).
The introduction of XR/Spatial Computing to the world at scale (both virtual reality and augmented reality) resembles more closely the original introductions of film (late 1800s), radio (early 1900s), and television (1920s/1930s). These devices changed the perception of reality for millions of people. The first radios were very much magical boxes where voices could be heard from anywhere in the world. They were exotic and mysterious. The first films shocked audiences. The advent of television in the home-where tiny people would beam into a glass tube box and entertain and inform you-this was all new and transformational. Complete ecosystems needed to be built to support these new mediums: television and radio stations, film studios, directors, producers, actors guilds, DJs, bands, record contracts, movie theatres, broadcast news, commercials/sponsors, and much more. These mediums touched people, they moved people, but none approach the closeness of interaction of XR.
XR devices are worn, and they interact with the visual and related sensory systems of the user in an intimate and deeply close way. SCUBA diving and wearing glasses (prescription, sunglasses) are the closest in form, but they are still pretty different things. For billions of people, XR is something truly new. Because XR connects so closely with the user, the requirements for greater and greater perfection are very high.
There is also no clear roadmap for how to build XR systems. The only way forward is to build something, get it into proof-of-concept and pilot sites, learn, iterate, design, build, test, and deploy again and again. The agile loop is a great methodology for software. It also works well for systems that have both complex software and complex hardware (like XR) — but full system agile loops are slower and much more expensive than software only agile loops. The pressure to execute complex system development in an agile format has led many of the top players in XR (including Magic Leap) to develop their own pilot factories and sophisticated, agile supply chain capabilities.
All of this effort is to gain unique and specific market knowledge and feedback, so that the efforts of a particular team and product direction may be accelerated by better understanding the early needs of customers, partners, and developers in the wild.
As an example, while we were developing the Magic Leap One (launched in August of 2018), we were also developing the Magic Leap Two (launching this year, 2022) as well as future generations beyond the Two. A number of the innovations in the Magic Leap Two (now seen in public) required significant research, investment, and testing (including the AMD chip). Practically all of the improvements were informed by internal testing as well as continuous feedback from early users, developers, and pilot customers. These cycles of learning are critical in any product development process, but are fundamentally key in something so new and disruptive as XR.
(3) XR needs 3–9 cycles of learning to move the market.
My initial thought was that 3 major cycles of learning (which I thought of as the Mercury, Gemini, and Apollo phases) would be required to move the market from tens of thousands of early enthusiasts and development partners to hundreds of millions of daily active users in both the enterprise and consumer sectors.
My current thinking has evolved in the following way: 3 major cycles of learning are critical to moving a major segment within one of the many emerging markets for XR, vs. the market moving forward as a whole (enterprise and consumer). The more refined and targeted your view of a segment of the market, the better. Within each of these segments, I see the need for 3 major cycles of learning, which can then cascade and propagate out into a related sector, triggering 3 new major cycles of learning.
In the enterprise XR sector we will likely see that a number of segments become commercially strong after 3 cycles of learning, but that almost every part of enterprise work has good adoption after 6 cycles. The XR consumer sector, specifically the mass adoption of consumer AR, will likely fully bloom after 9 cycles. Consumer AR will be deeply informed by the innovation and market success of really good enterprise AR (cycle 2, which we are in now, has great innovations coming, including the Magic Leap 2). This does not mean that consumer AR will not exist, but that it may take 9 full cycles of learning (including the critical 3 from enterprise) in order to become something billions of people use every day, all day.
(4) Pessimism or optimism?
Overall I remain very optimistic. The best teams in XR are capable of designing and delivering refined, beautiful, and functional systems that could be used all day, every day, by practically everyone. When I started in 2011 every single aspect of the system design was a major challenge, sometimes on the edge of impossible. That is no longer true. We will get to the size, weight, FOV, refined optics, tiny sensors, small form factor, battery power, CPU/GPU, etc. that we need. The sheer scale of investments by major players, coupled to ongoing innovations from adjacent industries (like mobile computing, robotics, and drones) will continue to drive the availability of the unique components and subsystems necessary to make great XR.
XR is a field that needs continued R&D progress towards an amazing long-term vision, significant and sustained capital, and support from all players in the ecosystem. The number of ongoing innovations and inventions required by the brilliant teams working in the field is hard to imagine (think many thousands of patents per top player in the field). The XR field is still so new that each day can be ripe for discovery, like wandering into a new, uninhabited land.
I tend to believe that AR and VR will synthesize into one well designed, sleek system, but that there will also be many flavors to accommodate the needs of many types of users and creators.
In its fully realized form, XR easily subsumes many previous media and computing devices: mobile phones, tablets, laptops, watches, and movie theaters. You will be able to conjur your own theme parks anywhere, and you will be able to project yourself into many forms (cartoony or photoreal) practically anywhere in the modern, connected world. XR is on a trajectory to change how we work, how we play, how we communicate, how we buy things, and how we create.
The field needs both patience (to achieve the broad visions) and urgency (to delivery ongoing improvements and value at increasingly faster rates). The social separation that was forced on us by the pandemic has also accelerated a new way of working: distributed, live anywhere, work virtually. XR (done well) will make this new form of how we work and live become a permanent part of life.
R. Abovitz
March 14th, 2022
Earth
