After Smartphones: Hold the Metaverse, the Cloud Assistant Era is Coming
SEE LAST PAGE OF THIS REPORT Paul Sagawa / Tejas Raut Dessai
FOR IMPORTANT DISCLOSURES 203.901.1633/.901.1634
psagawa@ / email@example.com
January 24, 2020
After Smartphones: Hold the Metaverse, the Cloud Assistant Era is Coming
The emergence of new device types (e.g. connected home, wearables, etc.), combined with improving communications networks and the inherent power/cost advantages of cloud datacenters will shift the nexus of consumer computing from the smartphone to AI rich, cloud-based platforms. The new paradigm will better support growing user needs for cross-venue portability (e.g. sessions that follow from device to device), new I/O modalities (e.g. voice, gestures, context, etc.), device independence (e.g. operation w/o the smartphone present), and inclusive ecosystems (e.g. open interfaces and interoperability). Smart home, infotainment, healthcare and fintech applications are the likely catalysts for a shift that will empower AI assistants while deemphasizing the smartphone OS. This is the direction charted by Google and Amazon, but largely counter to Apple’s market development roadmap.
- Competing visions: Star Trek or “The Metaverse”. The Star Trek computer, which infallibly answered queries and executed spoken commands while constantly monitoring the Enterprise, was an inspiration for many tech luminaries. Other industry execs have a near obsession with the sci-fi novel Snow Crash, which introduced the term “metaverse” to describe a virtual environment where humans can interact as electronic avatars. We believe that a coming consumer tech paradigm shift will take us much closer to the “Star Trek” vision, but that the “metaverse” concept faces technical and cultural obstacles that will prove unsurmountable for many years. This has real implications for companies that have set their development strategies from these ideas.
- Device proliferation, wireless ubiquity and cloud power. 3 major trends push us toward the “Star Trek” scenario. First, the cost/performance of cloud-based hyperscale datacenters is much better than alternatives and continue to improve. Second, wireless networks are getting faster, more available and less costly, from wide area (4G shifting to 5G), local area (WiFi plus new home networking standards), personal area (Bluetooth), and device2device transactions (NFC). Finally, new types of connected devices have gained traction, with smart home elements (e.g. smart speakers, connected appliances, etc.), wearables (e.g. smart watches, smart earbuds, etc.), public connections (e.g. touchless POS checkout, internet of things, etc.), and other types of devices (e.g. automobiles, gaming consoles, etc.) all emergent. We believe these trends will challenge the current smartphone dominated paradigm.
- These trends enable new use cases. The combination of new devices, better networks and hyperscale datacenters are combining to drive new applications. We see substantial opportunity in areas such as smart homes, health/wellness, payments/commerce, gaming/infotainment, and many others. For example, the newly formed CHIP alliance with AMZN, GOOGL, and AAPL will be a huge catalyst for smart home applications. Similarly, smart watches capturing health data to be analyzed in the cloud could catch problems earlier and aid in accurate diagnoses. There are many other exciting possibilities.
- New use cases place new demands on platforms. Today, the smartphone is the primary consumer computing platform, often acting as a hub for other devices. This paradigm faces challenges. Cross-venue portability: Users will want applications and data to follow them from device to device, such as transferring a streaming game from a portable device to a living room TV or expecting an office PC to remember a request made from a phone. New I/O modalities: In addition to text, clicks, and swipes on video screens, users may interact via voice commands, gestures, eye movements, or context etc. with devices that may not have a screen at all. Device independence: Users may want full interactions without the presence of another device (smartphone). This is obvious for fitness and healthcare applications, but reasonable in many other contexts as well. Inclusive ecosystems: As smart homes and IoT applications in the environment proliferate, consumers want confidence that everything can work together without committing to a single vendor’s solution for everything.
- The AI Assistant is the best bet for the new alpha platform. AI assistants are more than just voice commands. While speech recognition is an important modality, perfectly suited for hands-free (and screen free) interactions, gestures, eye movements, context, and other mechanisms will be alternatives, as will plain “text, swipe, click” on a touchscreen or keyboard. Assistants will handle all of it. They will also step in front of iOS or Android, managing access to apps and files, providing common utilities (e.g. search, map, calendar, messaging, media playback, payments, identity management, etc.), facilitating cross-app integration, and handling security. The AI assistant could act as a smart monitor across all the systems, flagging security issues and handling mundane requests without human involvement.
- GOOGL and AMZN have significant advantage. Alexa and Google Assistant have the lead in establishing the cloud-based platform paradigm, levering their extraordinary hyperscale datacenter infrastructure. AMZN is ahead at home, with a 66% share of the smart speaker installed base and the broadest roster of 3rd party partners. However, GOOGL has 2.5B Android devices in the field, strong loyalty to its franchise applications, and industry-best AI development chops. We believe that the next 5-10 years will see aggressive competition between the two to establish leadership across venues – the home, the car, the office, public spaces, and on the person.
- A boon to 3rd devices and software. Plug n play interfaces, useful APIs (e.g. authentication, payments, translation, etc.), cross-app integration support, and ubiquity will enable many new products to tap into the platform and reduce the time/cost to market. We see real opportunity for smart home devices – e.g. appliances, environmental controls, infotainment, security, and others – driving future upside for companies like Sonos, iRobot, Samsung, and many others. Omnipresence, flexible I/O and simple decision-making capability will enable a range of new application areas for thoughtful developers. Through this, AMZN and GOOGL will be looking to differentiate their ecosystems, via exclusive partnerships and acquisitions, adding potentially lucrative options for 3rd parties.
- Bad news for AAPL. Siri was first to market, but AAPL’s device centered philosophy contributed to a failure to push its further development, allowing GOOGL, AMZN and even MSFT to surpass its capabilities. Even now, AAPL seems to have underinvested in pursuing independence for its successful Watch and AirPod franchises, keeping them largely tethered and proprietary. Without the datacenter firepower to match Alexa and Google Assistant, and with an insistence on the iPhone as the master platform for its ecosystem, we believe AAPL is at risk of ceding the advantages of the cloud to its rivals.
Where We’re Going, We Don’t Need Smartphones
A lot of Silicon Valley’s illuminati look back on Star Trek’s omnipresent computer as a major inspiration. When Captain Kirk needed something, he wouldn’t need to pull out a smartphone, he just asked the computer a question or issued it a command. For other execs, Neal Stephenson’s 1992 sci-fi novel Snow Crash is the more compelling inspiration, describing a “metaverse” where humans interact as avatars in a virtual reality environment. We believe that the Star Trek computer is much closer than the metaverse.
Three factors have arisen to make it possible. First, the rise of hyperscale datacenters has dramatically reduced the cost of processing and storage, while raising the bar on performance. This coincided with the rise of hugely popular online franchises, like Google Search, Amazon Shopping and Facebook. Second, wireless connectivity is becoming faster, cheaper, increasingly reliable and more available. 5G for the wide area, WiFi and new “CHIP” (a fledgling coalition amongst AAPL, AMZN, GOOGL and the ZigBee alliance) protocols for homes and offices, Bluetooth for short range connections, and NFC for very close transfers will cover the waterfront with standards that will impel lower component costs and wider adoption. Finally, new device types – e.g. smart speakers, smart watches, smart earbuds, other wearables, connected appliances, and many others – able to access powerful datacenters in the cloud via ubiquitous networks at increasingly compelling price points are coming to market and catalyzing new use cases.
These applications present four key challenges for today’s smartphone centered paradigm. 1. More and more, users want profiles, permissions and applications that follow them from device to device and from venue to venue. 2. New device form factors ask for I/O alternatives to clicks, swipes and text on screens. Voice, gesture, eye movement, physical context, proximity to others, and many other things will be important inputs, while output may play to any screen, not just the one on the device. 3. A smartphone may not always be present. Users will want access to as much of their digital lives as possible on any device without having to tether to their phone. 4. As smart home devices and wearables expand their influence on daily life narrow ecosystems will feel increasingly limited.
We believe that the next alpha platform for consumers will be cloud based, able to coordinate across a universe of devices and fluent in a wide variety of interaction forms. It will be as ubiquitous as a user chooses and as convenient and powerful as possible. In short, we expect souped-up versions of today’s Alexa, Siri and Google Assistant, with powerful AI to anticipate needs, interpret ambiguous requests, make thoughtful recommendations, and simplify complexity. Meanwhile, the Snow Crash metaverse
A cloud-based AI assistant dominated world is obviously good for the companies with the best cloud-based AI assistants, that is, GOOGL, AMZN, Tencent and Baidu. It is good for the companies that make devices and applications that can tie to these open ecosystems – 3rd party wearables, connected home devices, venue bridging games, and so on. It is also good for the companies that make components that enable the broader experiences inherent in our scenario – communications interfaces, sensors, and datacenter piece parts. It would not be good for companies who are betting on the continued dominance of the smartphone vs. the cloud, that is, AAPL.
These Are the Voyages of the Starship Enterprise
Bill Gates, Larry Page and the late Steve Jobs have all mentioned a common inspiration – the computer on Star Trek. When Captain Kirk had a question that Spock couldn’t answer, or he wanted to issue a general command to the ship, he spoke aloud to the computer. Of course, Star Trek was created in the punch card era, and it was probably just as foreign to imagine a captain typing on a keyboard as it was to speak to the computer aloud, but to many of the early lights of Silicon Valley, that computer on the Enterprise was a revelation.
The idea was a natural interface that required no special syntax, that required no special input equipment, that could be operated from wherever you were without interrupting what you were otherwise doing. It interpreted your requests quickly and accurately then intuited what you needed and did it flawlessly. The inspiration of the Star Trek computer is obvious in Google Search and in Amazon’s Alexa. It is inherent in the initial conception of the iPhone, justly hailed for its revolutionary intuitive interface.
This paradigm remains a guiding light for many technologists (Exhibit 1), with AI the key tool to build it and Alphabet, and Amazon its leading evangelists.
Exh 1: A Timeline of Major Computing Eras
Welcome to the Metaverse
Neal Stephenson’s 1992 novel Snow Crash is credited for introducing the terms “metaverse” (a virtual reality space where people can interact with each other and with technology) and “avatar” (the digital representation that a user chooses to use within the metaverse). While Snow Crash owes its own debt to earlier cyberpunk writers like Philip K. Dick (who wrote the stories that inspired Total Recall, Blade Runner, and Minority Report) and William Gibson (who coined the term “Cyberspace”), it has been hugely influential, inspiring movies like Tron and The Matrix as well as a generation of Silicon Valley engineers. In the early days of Facebook, Mark Zuckerberg made the novel required reading for all new hires, while the builders of Microsoft’s Xbox Live cite it as the aspirational goal.
The metaverse of Snow Crash is a virtual reality, structured as a digital city with streets and buildings and such, where people can project their consciousness into digital avatars. In the metaverse, users as avatars can interact with one another, enjoy entertainment, execute transactions, and nearly any other activity in which they might engage in real life. Many technologists believe that there is a real-world race to build such a fully immersive environment, and that the winners could dominate a future electronic economy.
Faith in this paradigm seems inherent to the commitment that companies like Facebook, Apple and Microsoft have for virtual and augmented reality technology. Indeed, The Information reports that Apple CEO Tim Cook touted Augmented Reality as “the next big thing” while receiving an award in Dublin, Ireland this past week. If so, it would be a big step on the way to the metaverse, but we are skeptical.
Where are We Now?
Today, the smartphone, in the pockets of more than 3.5 billion people world-wide, is the dominant consumer platform (Exhibit 2). The smartphone operating system, be it iOS or Android, governs the way the platform is used, providing default utilities and serving as a gateway for downloaded apps running on the platform to access additional resources from the internet. Most other devices are subsidiary to the smartphone, taking instructions from it and depending on it for services and access to the cloud. Networks have been too unreliable, and often, too expensive for wearables and other secondary devices to rely on the cloud for processing and storage which are still considered too costly, bulky and power hungry to include.
Exh 2: Global Installed Base of Device Types in Comparison with Adult Population
Exh 3: Major Technology Shifts are Converging to alter Computing Paradigms
In this paradigm, the primary inputs are taps, swipes and awkward thumb typing on the smartphone touch screen, while output is largely the screen itself plus audio via integrated speakers or earbuds/headphones connected increasingly via Bluetooth. The user’s identity is intrinsic to the phone itself, which acts as a security screening mechanism for the apps that lie behind it. The smartphone is the digital you, offering myriad uses on-board and a hard gateway to services from the cloud beyond.
This is obviously a major step away from the Star Trek scenario and a giant leap from the metaverse (Exhibit 3). Captain Kirk didn’t tap his instructions into a hand-held device on the bridge of the Enterprise and the metaverse cannot be contained to a six-inch screen. If either of those aspirational paradigms is to take root, the smartphone will have to step aside as the alpha-dog element of digital life.
Changes are Afoot
We see three major technology trends that could challenge the current status quo. First, the computing and storage resources available from cloud-based hyperscale datacenters continue to expand at an extraordinary pace, while scale economies drive the cost of those resources down. This is particularly important for AI-based applications, such as natural language processing, image identification or recommendation engines, where specialized processors and massive databases can be invaluable, and cannot be matched by the computing and storage available on a smartphone.
Second, wireless communications networks are getting better – e.g. faster, lower latency, more available, more standard, and cheaper. 5G is an obvious element of this. Compared with 4G LTE, 5G networks will have much greater capacity in each cell, much lower latency, higher average and peak speeds, better potential coverage, and much lower cost per unit of capacity. At shorter distance, the WiFi 6 standard, just certified by the IEEE, will increase speeds, improve congestion, extend distances, and reduce power draw for connections. In concert, new protocols pending from the just inked “CHIP” pact between Google, Amazon, Apple and the ZigBee Alliance (which includes Samsung and the former Philips lighting amongst many others), will greatly reduce obstacles for the adoption of smart home products, which have been the
Exh 4: Global Smart Speaker Installed Base Forecast, 2018 – 2023E
Exh 5: US Smart Speaker Installed Base Share, 3Q19
Exh 6: Ownership of Smart Speakers per US Household
“next big thing” for nearly two decade amidst bitter battles between incompatible standards. Also forthcoming is a new standard for Bluetooth focused on improving audio performance with standardized codecs, providing precise location reckoning, supporting multicasting applications, extending its range and reducing its power requirements. Finally, Near Field Communications (NFC) chips have become extremely cheap and low power, providing an easy standard for exchanging data between devices at very close range (6-12 inches). This has proved very successful in facilitating mobile payments with potential for other applications.
Finally, new device categories have begun to gain traction (Exhibit 4, 5, 6). The first Amazon Echo hit the market in time for Christmas 2014, belatedly followed by Google’s Home in 2016. Today, more than 66 million US households have smart speakers of some ilk, with a total US installed base of more than 133 million units. Of these, a bit over 60% are Amazon Echoes, another 24% are Google Homes, with the remaining share split amongst a number of vendors, such as Samsung, Sonos and Apple. On a global basis, the installed base of smart speakers is well over 200 million units. These devices are not only untethered to a smartphone but can also act as an alternative control mechanism for connected household appliances, system controls (thermostats, lighting solutions, etc.), home security, and other smart home devices. Coming off of torrid 87% installed base growth in 2019, the smart speaker category is well placed to continue broad expansion. Video streaming devices, such as the Google Chromecast, Amazon Firestick or Roku Ultra, sell for less than $100 and offer easy connection from the internet to the living room TV. At the start of 2019, Amazon claimed an installed base of better than 30M for its Firesticks, topping Roku which had reported 27M active users. Extrapolating, we believe the total market installed base likely exceeds 100M.
Exh 7: Brief Overview of the newly released Wi-Fi 6 and Bluetooth LE standards
Exh 8: Wi-Fi 6 adds 1.4x more speeds and 4x more capacity for better connectivity
Apple’s efforts in the connected home have been tepid to date, but it has been the bellwether for wearable devices. The Apple Watch has been a been a hit, selling about 33 million units in 2019 up nearly 50% YoY and accounting for about half of the total smart watch market. FitBit, recently acquired by Alphabet, is the number two vendor with about half of Apple’s volume with Samsung, Huawei and a number of other brands filling out the list. Health and Fitness has been the killer app – the Apple Watch employs sensors to track a wearer’s heartbeat and other vital signs, along with logging mileage and altitude for the athletically inclined. Apple AirPods – Bluetooth connected earbuds able to capture spoken commands and relay them to the iPhone that is presumably in your pocket – have been another big hit, with an estimated 60 million pairs sold in 2019, double the volume shipped in 2018. As with Apple’s previous hit products, copycat smart earbuds from Amazon, Samsung, Google, and many others have followed AirPods to market, looking to address the huge Android installed base.
Together, new devices, better networks (Exhibit 7, 8) and powerful datacenters will enable extraordinary new use cases (Exhibit 9). For example: the smart home has been a “next big thing” for so long that it has become a bit of an industry joke. The big problem was turf wars over a grab bag of largely incompatible network schemes, all of which presupposed a single “hub” device that would coordinate the activities of the house. Of course, every company wanted to be the “hub” and none of them wanted to cede control to another company’s “hub”. In 2014, Alphabet’s Nest Labs proposed a new architecture called Thread that eliminated the need for a hub by interconnecting all devices into a non-hierarchical peer to peer network (Exhibit 10).
Exh 9: Converging technologies will enable extraordinary new use cases
Exh 10: Architecture of Google’s proposed Thread Protocol for IOT connectivity – which forms the basis of CHIP alliance architecture
This approach had several advantages – it had no single point of failure, it was designed to minimize power demands on member devices, and it was built atop IPv6 and 6LoWPAN, well established and open protocols. Thread will be the basis of the new CHIP alliance’s architecture, and with the support of Apple, Amazon, Google and the Zigbee Alliance (one of the previous grab bag standards), the rest of the industry will have little choice but to fall in behind the architecture, comforted by the lack of a dominant hub (Exhibit 11).
With this architecture, homeowners and developers can build future proof home nets, confident that most connected appliances, automated controls (e.g. thermostats, lighting, window shades, etc.), security devices (e.g. smart doorbells, nanny cams, motion sensors, smoke alarms, etc.), entertainment equipment, and other household electronics will be supported. Fanciful applications like firing up the coffee maker from bed or monitoring a robot lawn mower from the living room couch will be easy to set up and operate. The living room TV audio could be routed to smart speakers around the house to provide an ad hoc surround sound or to help you keep track of the game while you complete your household chores.
We expect that the number of networked devices in the typical home will multiply as the new architecture makes implementation and operation simple. Expect sales of wireless home routers supporting the new protocols and in smart speakers to serve as touchless control points to continue their surge. As consumers get accustomed to the convenience of having information about their homes at their fingertips and an easy mechanism to accomplish household tasks, expect meaningful new growth in spending on appliances and what all.
Exh 11: Overview of Zigbee CHIP project involving AMZN, GOOGL, AAPL
The 75 million Americans in the Boomer generation are facing the reality of old age – aches, pains, infirmity and death. Meanwhile, US spending on health care grew 4.6% in 2019, taking it to nearly 18% of GDP (Exhibit 12). With those aging boomers sure to keep the nation’s attention on improving healthcare outcomes, while the reality of healthcare costs remains a front-page topic, new tech may have the best answers.
The Apple Watch and competitors like FitBit today can monitor a wearer’s heartbeat and sleep cycles, tomorrow will track blood oxygen levels, and soon could capture further metrics like blood pressure or blood sugar. Smart watches also keep close count of physical activity – steps, miles, changes in altitude, and the like – for athletes and the health conscious alike. This data has enormous potential value for accelerating the diagnosis of many health problems and for monitoring treatment, as well as promoting fitness regimens and other preventative initiatives. At the same time, the data collected by these devices is also fuel for extremely powerful cloud-based AIs designed to improve diagnostic tools for doctors and to inform public health policy. Combining all of this with things like video telemedicine, home blood testing, and prescriptions by delivery, we believe tech advances can deliver on improving outcomes at lower costs.
Who am I?
Mobile payments are becoming a big thing – they already are in China, where WeChat Pay and AliPay have replaced cash and pre-empted cards at many merchants. ApplePay and GooglePay have gained considerable traction in the past two years, with well over half of US retail locations now accepting mobile payments and
Exh 12: Breakdown of US Healthcare spending in 2019 – large pockets could be addressable through assistant administered healthcare services
Exh 13: US Digital Wallet Transaction Volume and Growth, Forecast 2018 -2023E
Exh 15: Consumers consistently rate tech companies higher than banks on key functions of banking
an estimated $10B in transaction volumes on ApplePay in 2019. Still, the current process still requires the customer to take out their phone to connect to the POS terminal and complete the transaction.
Amazon would like to take that a step further (Exhibit 13, 14, 15). Its Amazon Go stores eliminate those POS terminals entirely. Customers check-in with a barcode on their phone when they walk in the store and their purchases are automatically registered without physical check out when they walk out the door. The key is using the phone as a secure element of identity for authentication. In the future, even the check-in could be automated, with an indicator at the door or a custom tone to confirm the shopper’s identity, which could be authenticated from a wearable or even ascertained by facial recognition.
In this scenario, the smartphone could be unnecessary. If secure identity could be established without one, wearables (or even your face) could be used for many applications beyond retail payments. Your watch could replace all sorts of tickets. It could be a universal ID badge, ring full of keys or even a driver’s license. We expect considerable development along these lines in the years to come.
The Games People Play
Games are a $150B market growing at a double-digit pace. 5G, with its extraordinarily low latency, has the potential to untether higher performance games that could be streamed from the cloud (Exhibit 16). We believe this has the potential to be a major catalyst for extending the reach of gaming to new users and new venues. (Game Streaming: Not Everyone Can Win) New devices will already play a role in this – smart speakers are already starting to tout trivia games that can be played at home and it is a matter of time before compelling games for watches or even smart earbuds emerge. At some point, and we are cautious of the technical and societal hurdles that must be resolved, augmented reality glasses will be realistic, with gaming likely to be the killer app.
Streaming audio and video have been obvious drivers of device proliferation. Streaming sticks, like the Roku Ultra, Amazon Firestick and Google Chromecast, exist only to facilitate the distribution of video to preexisting TVs, while streaming music has been the primary driver of adoption for smart earbuds like AirPods, for smart speakers, and second only to health and fitness as a key app for smart watches. We expect the wish for personalized entertainment wherever, whenever and however to be a key app for new devices and services going forward as well.
What’s the Catch?
Home automation, health/fitness, payments/security, and infotainment are not the only growing use cases pushing consumer technology forward, but they are compelling enough to raise dissatisfaction with the current smartphone dominated paradigm. We see four major challenges for the industry to address in the next few years (Exhibit 17, 18):
Cross Venue Portability – A family doesn’t want their home settings to be unavailable just because dad took his iPhone with him to work. At break time at work, a millennial might want to use their office computer to pick up a game at the same point that she was at on the console at home. A shopper would appreciate the ability to reference a product search made at home at a store kiosk. Kids watching a movie on a screen in the backseat of a car would love to finish it on the home TV. An avid runner would prefer to listen to his
Exh 16: Snapshot of Global Gaming Sales by Geographic Segments, 2019
music on a run or a swim without having to lug along a smartphone. Indeed, some users might like the option to easily swap one smartphone for another on a temporary basis. This flexibility is difficult to accomplish while a single device is designated as a master.
New I/O modalities – The smartphone touch screen, along with its language of clicks and swipes, was one of the major advances introduced by the original Apple iPhone. Since then, the vocabulary has only gotten richer and the performance of onscreen keyboards has continued to improve. Smartphone displays have also gotten bigger with time, the better to show high resolution images or videos. Still, there are limitations – for the most part, accurate input beyond single clicks on the right half of the screen requires both hands and the user has to be looking at the screen.
The smart speaker phenomenon has pushed forward spoken commands as an input and audio as an output, adding to the automobile where manual inputs and reading from a screen have always been problematic. We believe smart earbuds will further popularize this modality, which obviously requires no hands at all. However, this is not the only new I/O. Hand gestures, facial expressions, eye movements, the movement and orientation of the device, physical and social context, and other inputs could all be modalities that cue activity via various types of devices. Gamers, used to console remotes with a dozen or more buttons and joysticks, are particularly sensitive to the quality and flexibility of input mechanisms. Output may still be mostly visual, but sound will grow in importance and haptic feedback will likely become key in many contexts (Exhibit 19).
Exh 17: Mobile centered Design Paradigm is giving way for Ubiquitous Intelligence
Exh 18: Enablement Chains of the Mobile-App and AI Assistant Generations of Consumer Computing
Device independence – Having a single master device is limiting. First, in line with the cross-venue portability issue, sometimes users would rather not have their smartphone along. Second, the master device becomes a single point of failure – lose it or break it and suddenly, other devices and services don’t work properly in its absence. This is particularly bad for shared items, like smart locks or the TV remote. Third, it creates a specific security vulnerability – lose control of your smartphone and far more than your social media accounts could be at risk.
Inclusive ecosystems – The smartphone dominant paradigm creates lock in, not just for your choice of phone but for your choice of everything. The watch and earbuds (not to mention your chargers, connection cords, and other paraphernalia) you buy must be compatible with your smartphone, and in many cases, the brand preferred by the phone maker will have proprietary bells and whistles that cannot be duplicated by 3rd parties. It is not just connected peripherals – your choice of applications (e.g. messaging, maps, search, payment utilities, media services, etc.) are often dictated by your choice of a master device. Indeed, the manufacturer of that device will restrict your options for gaining access to those applications and take a cut of any revenue generated, raising the price to you. Your choice of other connected products, such as home appliances, home security and automation gear, gym memberships, health care plans, automobiles, and others, might be dictated by their appearance on a short list of favored vendors willing to license into the ecosystem surrounding your smartphone. Of course, smartphone vendors see this as the price a consumer must pay to participate in a well-curated ecosystem, but perhaps there are better ways to assure compatibility and to protect against 3rd parties with bad intent.
Exh 19: SSR Predictions for Capability Progression of Smart Assistants Software
A Cloud-centered Paradigm Can Bring us “Star Trek” Functionality
We believe the Captain Kirk computer is within reach. Shifting the nexus of control from each individual’s smartphone to a personal electronic assistant based in the cloud resolves most of the challenges presented by the evolving consumer electronics landscape. The assistant can follow from venue to venue with standardized security mechanisms established to make sure each user is who they say they are. All forms of I/O can be easily accommodated as appropriate to the venue and device, with AI embedded that can anticipate many needs without even requiring specific input. Such a platform would be obviously device independent and could be made highly inclusive via adherence to open standards and by making powerful APIs readily available to developers.
Running on a functionally limitless hyperscale platform, with access to mountains of data, backed up in multiple physical locations, and applying state of the art security, a cloud-based AI assistant would offer superior performance to a device based alternative at a much lower per user cost, presuming that wireless connectivity is sufficiently available. With broadly deployed 5G, we believe that presumption will be fulfilled.
In this context, we expect a few things (Exhibit 20). First, look for tethered devices, such as smart watches and earbuds to begin to untether, with integrated radios to connect directly to WiFi and 5G in the absence of a smartphone. Second, look for AI assistants to begin to step in front of smartphone operating systems to take user queries and commands directly to the cloud, similar in fashion to the way that WeChat fronts the OS for most Chinese consumers. Third, look for assistant platform operators, such as Google and Amazon, to aggressively court developers to build new cross platform applications on standard APIs. Fourth, look for devices to begin to support open home networking standards, to explore device to device connectivity without an intervening “hub” device and to promote accessibility via assistant platforms. We believe that we are beginning to see nascent activity along all of these lines.
Exh 20: Snapshot of Developmental Cycles in the AI Assistant Era
What About the Metaverse?
We believe that there are huge technical, business model, and societal issues with the idea of a metaverse that will take many years, perhaps decades to surmount. First, the concept is inherently immersive – deploying it on a smartphone screen, much less a smart watch or a home speaker is bound to be inadequate to the central idea. Proceeding from this, the preferred mechanisms for interacting within the paradigm, VR goggles or AR glasses, are seriously problematic. VR goggles entirely separate the user from “real” life, largely requiring the user remain safely stationary while experiencing the metaverse and shutting off interactions with other humans not similarly immersed. This will be VERY limiting (Exhibit 21). One of the triumphs of the smartphone era has been to take electronic life away from the desk and couch, dramatically increasing the time available for connected applications and integrating them into daily interactions. VR takes that back.
Meanwhile, mass market AR glasses – indistinguishable from ordinary glasses yet affordable – are not close to possible. We have written about this. (The Long Road to AR Glasses) Meanwhile, the social stigma that was quickly attached to the mundane Google Glass initiative was nothing compared to the cultural concerns that would arise from REAL AR glasses, able to react with information about objects in the field of view. Finally, the fictional metaverse was a shared space with participation from a wide set of commercial entities and nearly universal buy-in from citizens. On our current trajectory of incompatible VR and AR platforms, we would be more likely to get several highly fragmented metaverses, none of which would have the critical mass of either services or users to drive broad acceptance. Tim Cook may say that AR will be the next big thing, but it is not at all clear why it will be particularly useful to his assumed customers.
Exh 21: Summary of major hurdles for mass market adoption of VR/AR Hardware
Google and Amazon are on Track
If we are right, and AI assistants evolve to become the next alpha platform for consumers, it stands to reason that the companies with the best AI assistants now and the greatest commitment to moving them forward are the most likely winners. That is obviously Google with its Assistant and Amazon with Alexa. Each has a slightly different resume. Google, with its history in search has a towering competence in AI and has built that into Assistant. It performs better in understanding requests and responding to them and is often able to anticipate user needs ahead of queries. It has leading franchises in email, maps, video content, etc. and has built that expertise into its product. Google also has the advantage of placing Assistant as a default on most of the roughly billion Android smartphones sold each year (Exhibit 22).
Meanwhile, Amazon has a 60% share of the smart speaker market. By virtue of its head start in the market and Amazon’s aggressive courting of 3rd party partners, Alexa has been implemented as default on a wide range of other smart devices for the home able to take spoken commands and in many automobiles. While its lead has been shrinking, Alexa also has a larger ecosystem of home devices that it can control. With its dominant e-commerce franchise, Amazon is able to push its products to market more readily than its competitors and ordering via Alexa is a strong differentiating app. We believe that both companies are likely to lead the cloud assistant era (Exhibit 23).
Exh 22: Key Smart-home categories are set for explosive growth in next 5 years
Exh 23: SSR Summary of Comparison Between Top Digital Assistants
With both Amazon and Google now supporting the same control protocols via the fledgling CHIP alliance, we also expect the market to be friendlier to connected devices for the home. With development risks lessoned and scale economics likely kicking in for components, better products will become available and consumers and home builders will be more enthusiastic buyers. We see companies like iRobot, Sonos, Samsung, Signify (the former Philips lighting business), and others seeing success. There should also be opportunity for 3rd party wearables – companies like Jabra, Logitech, and others come to mind – and for new software and services companies that address the AI assistant platforms with cross-platform applications.
Apple and Facebook Could Be Waiting for the Metaverse for a Long Time
Apple was first to market with many of the innovations that are helping to catalyze the rise of the AI assistant, including the pioneering Siri, introduced more than two years before Alexa. It missed the boat on smart speakers but leads the regatta with its Apple Watch and AirPods devices. Still we are bearish on Apple’s positioning for the coming platform paradigm shift (Exhibit 24). Unlike Amazon or Alphabet, Apple is distrustful of the cloud, maintaining a laser focus on the iPhone’s role as the hub of its multi-device product strategy. As such, it has underinvested in Siri and kept her limited to the device – there is a different Siri on each Apple device you might own, and they don’t talk to each other very much. The vocal global minority that obsesses on data privacy (N.B. we don’t think most people care all that much, otherwise there wouldn’t be 2.5 billion people on Facebook) has given new conviction to Apple’s monomania about the primacy the smartphone. Given CEO Cook’s comments on AR as “the next big thing”, we don’t expect a commitment from Apple to a cloud-based AI platform as the center of the new consumer computing paradigm anytime soon.
Facebook doesn’t seem to be looking in that direction either. Facebook made a half-hearted attempt to front Android with a customized launch page that flopped and launched its Portal home video conferencing platform (which also looks like a flop), and its first take on an AI assistant, the tersely named M, has come and gone. The company has confirmed that it is working on another AI assistant, this time to work with Portal, but it seems a little bit underwhelming given the enormous lead that Google, and Amazon have established. We suspect that Facebook’s foot dragging here is related to Zuckerberg’s reported obsession with the metaverse and the strategic importance that he places on the company’s Oculus VR platform (Exhibit 25). If the Metaverse is the next big thing, then universal talking computer access is a distraction.
Exh 24: Highlight of mounting threats to Apple’s walled-garden approach
Exh 25: FB acquisitions in the last 2 years – spent nearly a billion dollars on acquiring Cloud Gaming and VR assets
Exh 26: Summary of Winners and Losers for Digital Assistant Paradigm Shift