Smart Homes: Are We There Yet?
SEE LAST PAGE OF THIS REPORT Paul Sagawa / Tejas Raut Dessai
FOR IMPORTANT DISCLOSURES 203.901.1633/.901.1634
psagawa@ / firstname.lastname@example.org
February 18, 2020
Smart Homes: Are We There Yet?
The smart home has been the “next big thing” in tech for about 20 years, but incompatible competing standards and awkward solutions have largely squelched adoption. We believe that these obstacles are finally clearing, opening the way for tech that will change the way consumers live. The biggest catalyst is the “CHIP” pact amongst AAPL, AMZN, GOOGL and the ZigBee Alliance that will have the dominant consumer tech companies supporting a common non-hierarchical, IP addressable networking standard that eliminates the need for a master hub device. We expect industry-wide transition to the new standard, driving down component costs, assuring interoperability, and dovetailing with performance improvements to complementary WiFi and Bluetooth networks. The second is the rise of AI assistants that can interpret spoken commands and smart speakers as access points, simplifying controls and damping the learning curve. We expect AAPL, AMZN and GOOGL to articulate much more compelling home networking stories, for 3rd parties to enthusiastically support them with new products, and for consumers and homebuilders to embrace the emerging paradigm. We believe AMZN and GOOGL will gain the most from this, with 3rd party participants, like SONO and IRBT, benefitting.
- The next big thing. Demos showing door locks, environmental controls, appliances, entertainment systems, and other consumer devices acting in concert automatically and/or in response to spoken commands has been a fascination for decades, with various tech companies showing their visions in elaborate ersatz houses built on the show floor at CES over the past two decades. Still, the reality has been “just around the corner” the whole time, as vendors pushed at least 6 different incompatible networking standards and pushed to claim the dominant role as the hub of it all. Consumers saw the tower of Babel and shrugged, putting it all off until someone made it easy and affordable.
- CHIP breaks the standards logjam. 2 months ago, GOOGL, AMZN, AAPL and the ZigBee Alliance – a group of companies supporting (but not necessarily exclusively) the most popular of the 6 disparate home networking standards – agreed to establish a common architecture for the home based on THREAD, an open, non-hierarchical IP-based, protocol originally developed by GOOGL’s Nest business unit, implemented by key silicon partners and recently supported by both AAPL and AMZN. The CHIP standard should have the advantages of inexpensive components, low power operation, and standard IP addressability, without elevating any one device to the role of “master”. We expect accelerated development by vendors and deployment by homeowners and builders, driving to near universal adoption.
- Advances in WiFi/Bluetooth will complement CHIP. THREAD is a low-power, moderate speed, resilient network intended to handle device to device communications, complementing WiFi (high speed network for tetherless user connections to the internet within a ~150ft radius) and Bluetooth (highspeed/low power network to facilitate transfers/streaming over shorter distances). New releases of these standards are due this year. WiFi 6 will increase connection speeds, reduce congestion from multiple devices and better manage power use. Bluetooth enhancements will add precise device location mapping and improved audio performance, including multicasting to many devices at once.
- Home speakers and AI assistants. In the 5 years since AMZN introduced the Echo and its assistant, Alexa, the global installed base of smart home speakers has grown to 200M+, used by more than 40% of US households. These devices, still dominated by Echo but with GOOGL’s Home now a strong #2, are primarily used to access information and play music/video but are well positioned as access points to manage networked home devices. This differs from the idea of a “hub”, as other devices can equally serve as access points (and an Echo can be easily swapped for a Home), but the convenience of voice control is undeniable. Behind the speakers, we believe the AI assistants – e.g. Alexa and Google Assistant – hosted in the cloud will become the organizing mechanism for many consumers, anticipating requests based on context and automating complex actions without tedious set-up.
- The things you could do. A common standard, reliable connectivity, simple voice interfaces and powerful cloud AI backing it all enable many use cases in the home. Automation is an obvious category, executing tasks – e.g. vacuum the living room, open the shades, dim the lights, start the coffee, set the thermostat, etc. – on command or automatically as part of an established routine. Monitoring solutions – for security, safety, environmental management (e.g. power efficiency, utility usage, etc.), etc. – will improve, fall in cost, get easier to use, and trigger automated actions as needed. Entertainment will be enhanced, with services decoupled from specific devices creating flexibility and convenience (e.g. music that follows you from room to room. Computing will be ambient, taking queries and commands wherever, whenever.
- AMZN, GOOGL and AAPL will take the lead. While the CHIP architecture eliminates the need for a designated hub to control a home network, we expect software platforms to facilitate implementation and use for consumers. For AMZN and GOOGL, their cloud-based AI assistants will likely play this role, while AAPL is more likely to focus on a dedicated default app integrated into iOS. The cloud solutions are likely to be more robust, enabling control through more device types, coordination with environments outside the home, and facilitating multi-person operation. 3rd party home tech companies will fall in line, adopting the networking standard and following APIs to make their products controllable by the platform applications. This will make automating a home much easier and future proof solutions for future CHIP upgrades.
- Winners and losers. We don’t necessarily see losers here. Accelerating adoption of networked smart home solutions opens a myriad of opportunities for platforms, device makers and software developers alike. However, we see AMZN and GOOGL as particularly advantaged with their leadership in smart speakers and AI assistants. Forward thinking device makers, like IRBT, SONO, Signify, Samsung, LG, TP-Link, Ecobee, NTGR, LOGI, Dyson, etc., and THREAD chip suppliers, like NXP, SLAB, Freescale, and QCOM, also stand to benefit as buying criteria shift toward networked solutions. Note that adoption of the new standards will likely take several quarters to build to volume.
We’ve all seen the demos. From the animatronic family in Disney World’s “Carrousel of Progress” to the ersatz houses built right on the show floor at the Las Vegas Convention Center for CES, the smart home has been the “next big thing” for decades. The lights dim, coordinated with the shades dropping and the music starting, automatically setting the evening mood. The thermostat adjusts the temperature, the oven automatically preheats, and the doorbell sends video of guests waiting on the stoop to the kitchen display which had been pacing you through a dinner recipe. The kids move to the den and their video game follows them so they can pick up in the exact same point in the new venue. It’s the same demo. The brands may have changed, but the vision is reliably consistent. And still in the future after all these years.
This time, the future may be upon us. For years, the biggest obstacle to smart home solutions has been a protracted tug-o-war over standards. As many as 6 different wireless networking schemes were in play and most required a single device to act as a hub to connect and command the rest of the home. Needless to say, every vendor wanted to be the hub, and most made it difficult to connect anything that wasn’t explicitly signed on to the ecosystem. Home lighting existed on a separate system from connected thermostats which were didn’t play on the same system as home security systems which were entirely distinct from entertainment.
In December, AMZN, GOOGL, AAPL and an organization called the ZigBee alliance that had backed the most popular networking standard, and that included most of the leading chipmakers for consumer devices announced a new pact called “Connected Home over IP” or CHIP with the intent to establish a standard specification as a fair playing field for the industry. At its center is the THREAD protocol, a low power, non-hierarchical (all connected devices can communicate with one another as peers) solution based on IP, which enables the network and its devices to be addressed directly from the internet. There is no need for a single hub – conceivably any device could act as a control point, although it is likely that users will migrate to favored methods of interaction. This is a big deal – standardization means volume, volume means lower costs and market momentum.
In addition to this, both WiFi and Bluetooth have new versions hitting the market this year, dovetailing with the THREAD control network to deliver high bandwidth services to users. WiFi 6 will raise speeds, increase range, improve reception, and enable many more devices connected to each router. Bluetooth 5.1 specifies a standard codec for reliable interoperability, adds precise location reckoning, and supports multicast to several devices at once. These benefits will be added inducement for consumers to refit their homes.
Finally, we believe smart speakers (now in 40%+ of US homes) and their cloud-based assistants will become the primary UI for accessing the capabilities of the smart home, powerful tools for managing devices and tasks, and key resources that can intuit user needs based on context. This would put AMZN and GOOGL squarely in the driver’s seat. Other winners would be various connected device makers from appliances to security systems to home entertainment products, and the chip makers that will enable them to be connected. While AAPL is a founding member of CHIP and boasts a formidable base of loyal customers, we are concerned that their architectural focus on the smart phone as the primary organizing mechanism for smart home functionality is flawed.
The Tower of Babel
Sci-fi writers have long had a fascination with home automation. With today’s eyes, it can seem nonsensical to build at robotic housekeeper that does the dishes in the sink, but that’s what Rosie did on the Jetsons. Today, we have appliances like microwave ovens and immersion circulators that would never have occurred to Isaac Azimov back in the day, but Roombas and Hue lighting systems would be at home in the homes of the Brave New World.
Where we have fallen down most against the visions of the future from the past is in control systems. These visionaries would be appalled by the motley collection of remotes cluttering the typical American coffee table. The idea that setting the mood for a romantic dinner would involve separate systems to set the lights, draw the drapes, adjust the room temperature, and start the gas fire would seem silly. That a homeowner might have to haul a ladder into the bedroom at 3AM to shut down the infernal chirping of a smoke detector low battery alarm would be beyond comical. But that’s the world in which most of us live.
Technologists have been trying to solve the problem for decades and working demos of connected smart homes have been a staple of tech trade shows since before Y2K, but none of it has much traction. The biggest hold up has been the rivalry over who gets to control the action. Those millennial fights were over which device would serve as the ultimate hub of the connected home – the PC or the Set Top Box. A decade later Apple jumped in with its own self-serving candidate, the iPhone, while other home products went another way entirely with specialized hubs just for a single use case – a hub for lighting, a hub for home security, a hub for environmental controls, and so on. At least six different network standards were proposed and adopted by the various industry factions, not one compatible with the others. A choice on any of it would lock a homeowner to a single narrow ecosystem for that part of the equation with grave concerns as to whether other use case specific solutions would play nice in concert. No wonder the state of the home of the future is so dismal (Exhibit 1, 2).
Exh 1: Key factors causing home-networking protocol logjam for the past years
Exh 2: Overview of various wireless protocols currently used for home networking
A New Hope
In December 2019, Amazon, Apple, and Google joined forces to form the “Connected Home over IP” or “CHIP” group, pulling in the ZigBee Alliance (an industry group supporting the most successful older home networking standard including many key communications semiconductor manufacturers) to support what is planned to be a single standard (non-ZigBee) architecture for smart homes. The network specification will be built atop the THREAD protocol developed by Alphabet’s Nest business unit, released as an open standard in 2015 and gaining support from important players like Amazon, Apple, Qualcomm, NXP, Freescale Semiconductors and Silicon Labs in the intervening years. Unlike ZigBee, and most of the previously proposed standards, THREAD does not require a single master device to act as a hub for all the others (Exhibit 3, 4). It is non-hierarchical and any device can connect directly to any other device on the network. It is also based on IPv6, enabling any device on the network to be addressed from the internet – presuming the request is appropriately secured. THREAD is also designed to be very low power, so that battery powered devices (those pesky smoke detectors!) experience no extraordinary drain.
Of course, CHIP is not just THREAD. The CHIP control protocols will be designed to operate over any IP-based network, meaning WiFi and Bluetooth will be part of the rubric, complementing THREAD by enabling much higher speed connections and much greater compatibility with existing devices in the home. CHIP will work to avoid cross network interference between the underlying wireless networks. Automation,
Exh 3: Summary of Catalysts Driving Innovation in Home Networking
Exh 4: Overview of the new CHIP Project involving AMZN, GOOGL, AAPL
monitoring, entertainment and computing all on the same networks following the same control protocols! To quote Ghostbusters – Dogs and cats living together! Mass Hysteria!
This is all a very big deal. The leadership of Google, Apple and Amazon give it substantial oomph – smaller companies will have little choice but to support it (Exhibit 5). Unanimity will drive volumes, which will lower component costs for necessary radios and processors, which will drive further adoption, and so on.
It’s Getting Better All the Time
The formation of the CHIP working group dovetails nicely with advances that are coming from the WiFi and Bluetooth standards bodies (Exhibit 6). WiFi 6 (the first move was to finally ditch the complicated 802.11 naming conventions) is a big step. Using many of the technologies behind 5G wireless (OFDMA multiplexing, MIMO, and beamforming), WiFi 6 increases the capacity of a single router by a factor of more than 4x while raising the ceiling for connection speeds to 10Gbps. It improves the effective range, including reducing the degradation from walls and other physical objects. It also reduces the power drain on connected devices (Exhibit 7). Meanwhile, Bluetooth 5.1 largely focuses on the protocol’s performance for audio applications, increasing the maximum speed to 50Mbps, lengthening the range to 800ft and introducing multicast capability (multiple devices can tap into the same signal). It also adds precise location reckoning, improving its utility for finding lost things or fixing a specific placement (Exhibit 8).
Devices with need of high-speed connections for audio, video and other media types will see fewer dead spots and more consistent performance within homes (and elsewhere) while control traffic shifts to the purpose-built THREAD net (Exhibit 9, 10).
Exh 6: A Brief Overview of the inter-operable THREAD Protocol
Exh 7: Wi-Fi 6 adds 1.4x more speeds and 4x more capacity for better connectivity
Importantly, the new standards are backward compatible, keeping older devices functional while adding performance and enhanced capabilities to new ones
Exh 8: Brief Overview of the newly released Wi-Fi 6 and Bluetooth LE standards
Exh 9: Highlight of symbol characteristics for WiFi-6 waveform
Exh 10: End-device limits for THREAD and other comparable protocols – THREAD offers considerable improvements
None of the CHIP protocols require a single device to be designated as a control hub. Ambitious device makers can adopt the standard without necessarily ceding a dominant position to a would-be competitor. Still, we don’t see the home of the future as some sort of wild west of compatible yet completely independent devices. Something will be the primary user interface, and something will organize the disparate items of electronics into a coherent system (Exhibit 11).
We think those things will be voice commands and AI assistants (Exhibit 12). The world-wide installed base of smart home speakers has expanded dramatically, doubling in 2019 to more than 200M units (Exhibit 13). In the US, they are in use in more than 40% of households, answering queries, playing media and responding to commands. Some of these commands are to control smart home connected devices – “Alexa dim the living room lights” or “Hey Google, set the kitchen temperature to 70 degrees”. With the massive set of commands submitted to Amazon and Google and their substantial investment in the AI needed to process them, the voice recognition need to recognize commands is getting more precise and more reliable, even amidst background noise. With CHIP, setting up these speakers to execute instructions is going to get easier and support for them by the devices we buy will become nearly ubiquitous.
The AI assistants inherent in smart speakers will also enable increasingly complex requests – “Hey Google, set up the whole house in the same way as we did for the dinner party a month ago” or “Alexa, make a pot of coffee as soon as I wake up” should reliably result in a soothing atmosphere for your guests as they arrive and a fresh cup of joe in the morning. Eventually, the assistants will learn enough about you to automatically anticipate your requests – noting the evite for your party and setting the house without needing to be asked
Exh 11: Device centered Design-Paradigm is giving way for Ubiquitous Intelligence
Exh 12: Enablement Chains of the Mobile-App and AI Assistant Generations of Consumer Computing
or making that coffee every morning before calling you an Uber to take you to work. That all of this will be accessible via basic spoken commands in umpteen different languages removes significant hurdles to adoption. For example, in our experience, elderly relatives have a much easier time dealing with a voice assistant than a smartphone.
We note that the non-hierarchical nature of the pending CHIP architecture means that households are not locked in to a particular brand of speaker or AI assistant – you could plug an Echo into an otherwise Google Home dwelling and begin setting it up to control things without reconfiguring anything else. Of course, the new device will have much less data about you (e.g. preferences, schedules, vocal tics, etc.) so the platform stickiness will have to do with that rather than participation in a closed hardware architecture. Even without stickiness, removing obstacles to deploying smart home networking can only be a further boon to the already rapidly growing smart speaker and AI assistant market (Exhibit 14, 15).
What Do I Get from All This?
We see use cases in the connected home falling into four main buckets: Automation, Monitoring, Entertainment, and Ambient Computing:
Automation- Humans have long looked to machines to remove drudgery from their daily lives. Many appliances, like dishwashers and coffee makers, could be described as highly specialized robots that automate time consuming tasks that might otherwise be performed by hand. The home of the future will have more of that and even the job of setting up and initiating the task might be automated. Lighting systems, particularly LED lighting that can be adjusted by color palette, can be precisely calibrated to personal preferences by time and context. Moreover, multiple tasks to be performed by multiple devices in the home will be coordinated – e.g. electric shades drawn, lighting set, gas fireplace lit, and thermostat adjusted in a single command.
Monitoring – The smart home will automatically read utility meters both to report back to the service provider, but also to allow homeowners to manage their own use. The AI assistant can alert users to high-usage activities and devices or automatically adjust settings to economize when occupants are away. Security will be a big application – coordinating doorbells, nanny cams, motion sensors, smoke detectors and other sources to identify threats, notify users and take necessary actions (e.g. call the police or fire department). It can also monitor health care inputs and act as a lifeline for the disabled.
Entertainment – A smart home can help to eliminate remote control clutter and act as a single control point for multiple devices. Content can follow a user as they move around the house and be delivered with precision to any appropriate connected endpoint – the audio from a TV show can follow on speakers down the hall, and the video stream transferred to the bedroom TV. Kids can play the same game from different rooms or pick up the game from the same point in on a different TV if the parents want to use the living room for something else.
Ambient Computing – Make an appointment. Place an order. Ask a question. Call a car. Dictate and send a message. Check game scores. Get news updates and reminders. Many of the things for which you might pull out your smart phone or go to your PC could be accomplished with spoken queries and output sent to the most convenient device. The home of the future could be a lot like the Starship Enterprise where Captain Kirk simply makes requests aloud to a computer that executes them quickly and accurately.
Amazon, Google, Apple and Facebook.
Taming the tower of babel will be an unequivocal good for the big consumer internet platforms. The biggest beneficiary may be Amazon, which has lacked primary agency with its customers. By this, we mean
Exh 13: Global Smart Home Device Installed Base Forecast by Type, 2019 – 2023E
Exh 14: US Smart Home Revenue and Growth Forecasts, 2019 – 2024E
Exh 15: Breakdown of US Smart Home Revenue by Category
shoppers come to Amazon via a browser or smartphone app, clicking through from a platform controlled by someone else. Alexa in the home eliminates this previously necessary middleman, gives Amazon privileged access to the search that leads to the sale and lets Amazon play gatekeeper for other 3rd party services that might be accessed via Alexa. For example, Uber might pay a commission to Amazon to be the default ride service if an Alexa user asks for “a car”. With a 61% share of the global smart speaker installed base and with generally high marks for Alexa, Amazon is in the lead amongst would-be connected home platforms, one that could help the company bridge into broader use outside the home – in autos, in public forums, or even on a smartphone.
Google’s line of prosaically named Home speakers are number two in the global market and gaining share. The eponymous Google Assistant has the further advantage of the more than 2.5 billion smartphones using the Android operating system (Exhibit 16). Many of these devices have or will have Assistant as a default, giving Google another bridge from the outside world into the home. With its legacy as a dominant search platform and leadership in the science of AI, Google Assistant is consistently rated the best performing solution for its exceptional voice recognition, flexibility in executing commands, and, of course, its prowess at search. It also enjoys close links to Google’s other internet franchises – beyond Android and Search, Chrome, Maps, YouTube, Gmail, Drive, the Play Store, and Photos can all boast of more than 1 billion users. These products can be happily integrated as defaults in Google Assistant.
Exh 16: Active User Base Estimates for Popular Internet Platforms
Apple’s home strategy is a bit askew from Amazon and Google, relying much more on its iOS platform running on iPhones and iPads than on its Siri voice assistant. This creates some awkwardness, requiring a device in your hand to control the home and complicating the process for managing multiple users. Apple may find loyal iPhone users turning to Echoes or (gasp!) Google Home products to gain voice control, and in doing so, shifting platform dominance to an Apple rival, at least in the home. As all of the big internet platform players look to assert their dominion, ceding advantage in the connected home could be a big risk for Apple. Yet Apple isn’t really investing toward the voice control/AI assistant narrative, launching a poorly received home speaker in 2018 and allowing Siri to continue to choke on Alexa and Google Assistant’s dust. Part of this is Apple’s conviction that AR glasses will be the next big thing in tech, a vision of which we are quite skeptical (Quick Thoughts: The Long Road to AR Glasses) and part of it is Apple’s distain for the cloud. No matter, we believe that the company’s faith in the central role of personal devices is misplaced, at least in the home.
Finally, Facebook. Facebook has more users than anybody, counting nearly 2.5B for its main app, with the WhatsApp messaging platform pushing 2B on its own and Instagram topping 1B. There is plenty of overlap, but that is a LOT of people using Facebook services on a frequent basis. Still, however successful Facebook apps have been, CEO Zuckerberg has platform envy. Google has Android and those 2.5 billion users. Apple has iOS with over a billion users of its own dancing to its proprietary music. Amazon is building a platform in the home via those Echoes and Alexa. What does Facebook have? Less than 3 million Oculus VR gaming platform users. Like Apple, Facebook sees VR/AR as the next big thing and has invested accordingly. Its Portal video-conferencing focused smart hub for the home has been a significant failure. Recent rumors suggest the company is looking jump into the AI assistant race to counter the influence of Google Assistant and Alexa. We are concerned that it is just too late to catch up.
Who Else Wins?
We believe the combination of the CHIP standard, better performing networks and the rise of smart speakers/AI assistants as powerful and intuitive user interfaces will accelerate adoption of smart, connected home solutions. Those devices designed to the architecture have substantial new opportunity. We can categorize 3rd party devices by application type (Exhibit 17). Automation – smart lighting solutions from vendors like Signify and Samsung, appliance makers (Samsung, LG, iRobot, Dyson and others), smart thermostats (Google Nest, Ecobee etc.), nanny cams (Logitech, Amazon, ADT) Monitoring – security systems (Vivint, ADT, Honeywell, etc.), Smoke/CO2 detectors (Nest, Ecobee, etc.), Doorbells/locks (Amazon Ring, August, Google Nest, Yale, etc.) Entertainment – Smart speakers (Sonus, Amazon, Google), SVOD TV sticks (Amazon Fire, Google Chromecast, Roku) Ambient Computing Smart Speakers (Amazon and Google), On-line services. Home Networking Infrastructure – (NetGear, TP-Link, Google, Belkin) Components – (NXP, SLAB, Freescale, QCOM, etc.)
Please note that we do not expect the smart home market to take off immediately. This is likely to ramp slowly, with the best opportunities for investors likely more than a year away. Still, we believe that this could be a significant driver of consumer tech spending for a very long time
Exh 17: Summary of Winners for Home Networking