
I’m getting close to admitting defeat. And it hurts. I’ve long been a proponent of smart home hubs with multiple radios, particular those for Zigbee and Z-Wave, but it seems like the hub market is happy to ignore those in favor of Wi-Fi and Bluetooth only. And by “hub market”, I really mean the “big three” players who joined the game relatively late: Amazon, Apple, and Google.
Yes, there are still companies that support all four radios, which provides the widest breadth of compatible, energy-efficient products. Samsung’s SmartThings platform doesn’t seem to be going away anytime soon, for example. Wink is still around although, for some reason, their Hub2 hasn’t been available on many retail sites for some time. And there are some others such as Fibaro’s Z-Wave hub and Habitat’s Elevation hub, which includes the four major radios, and also includes Lutron support.
However, the “big three” have not only moved the smart home interface from phones to speakers, but they’ve also usurped smart home functionality by adding routines and a little programmability. As I noted earlier this year, these abilities to consolidate commands or automate your home is one of the key strengths of a true hub compared to a voice-activated device aggregator. And I still believe a traditional hub like SmartThings or Wink are the better solutions. I’m not sure the market agrees.
Why do I think that? Well, I haven’t seen any sales figures for the traditional hubs. But I have seen estimates of smart speaker sales. Back in August, it was reported that Amazon Echo and Google Home devices had a 50 million install base within the U.S. Apple’s figures are harder to pin down but keep in mind that every AppleTV or iPad can be a HomeKit hub and that both Siri and Apple’s Home app is available on hundreds of millions of IoS devices.
I suspect the install base of traditional smart home hubs is dwarfed by the Amazon, Apple and Google numbers. And aside from the Amazon Echo Plus devices, which do have a Zigbee radio, all of the other smart speakers and voice assistants from these three companies natively work with Wi-Fi and Bluetooth. Yes, adding Zigbee and Z-Wave devices to these can be done if the product supports a bridge to read in these radio signals and convert the data to Wi-Fi or Bluetooth.
Does the mainstream market care enough to do that in meaningful numbers? It doesn’t look like it. Instead, we smart home die-hards that want total control and a wide range of products are the only ones that care about radios whose name starts with a “Z”.
It doesn’t appear this will change any time soon either because Apple, Amazon, and Google are taking a methodical approach.
After a few years of HomeKit stagnancy limited to simple device controls, for example, Apple added automation support in iOS. Amazon and Google have more recently added support for Routines and Actions, which are more like a series of controls rather than automation but you can have smart devices do certain things based on time settings with an Echo Routine. And I wouldn’t be surprised at all if both of these companies add cloud-based automations to their products in the near future.
Personally, I’d rather have localized automations, but again, will mainstream consumers? Frankly, a little more storage and processing power in future Amazon Echo and Google Home products could mitigate that anyway.
Here are another few data points that suggest to me that quad-radio, full-featured hubs are on the endangered species list. In its quarterly results last week, Amazon said “The number of Alexa-compatible smart home devices has quintupled year to date to more than 20,000 devices from over 3,500 brands.

And when Google introduced its new Google Home Hub earlier this month, it noted that its Assistant now works with “ over 10,000 smart devices from over 1,000 brands.” Apple’s HomeKit is more limited, of course, but do you think SmartThings, Wink and other hub makers can say the same about their device and brand support?
Listen, I know this idea of doom and gloom won’t be well received by the passionate smart home readership here. And trust me, I’m not happy to share this opinion. Luckily, we can swim against the current and use our more functional hubs if we want to.
But the reality is starting to set in from my perspective: “Normals” care less about automation right now and are excited just to control their smart devices through a speaker or touch display. By the time they’re ready for a truly smart home with automations, the non-hub makers will have cloud services to provide it. And by focusing more on the user interface and brand integrations, the “big three” can always add more radio support in the future if the market demands it.
The end of the hub was precipitated by the Espressif ESP32. That’s a single SOC chip with wifi/BT for $2. It is a double whammy, the ESP32 is cheaper than Zwave/Zigbee and you don’t need a hub. ESP32 based devices are all over Amazon now advertising Alexa compatibility.
https://www.espressif.com/en/products/hardware/esp32/overview
Also note that AWS Greengrass has local control ability. “AWS Greengrass devices can act locally on the data they generate so they can respond quickly to local events, while still using the cloud for management, analytics, and durable storage. The local resource access feature allows Lambda functions deployed on Greengrass Core devices to use local device resources like cameras, serial ports, or GPUs so that device applications can quickly access and process local data.” This model lets you use the cloud for programming and setup, but then the actions execute locally.
These two items have combined to kill the old style smart home hub.
I mean I got a Lenovo display and it’s actually awesome. My housing environment literally changed cause I connected all my other devices to my Lenovo display, including my coffee machine. And to be honest, compared to other devices for the same price I got a huge screen, awesome speakers and so many useful features.
One of the most interesting recent moves was Amazon’s decision to include a zigbee radio in the second generation Amazon Show. A lot of people will now have a hub who don’t even know it.
And the main change is that these people will be able to add battery-operated devices like locks and sensors which are not typically available with Wi-Fi because of too short battery life, and not typically available with Bluetooth except for HomeKit because of the need to have a phone nearby.
They won’t be able to do as much with these devices as they could with a full featured hub, but it does fill the gap for android users. (iOS users already have the same options with HomeKit because of another stealth hub: the Apple TV.)
I haven’t seen anything official, But my guess would be that Amazon chose zigbee over Zwave because of the “simple set up” option which allows voice discovery of those zigbee locks and sensors. Z wave doesn’t offer that yet. Or maybe they just didn’t want to go with a radio option which can only be sourced from a single chip vendor.
In any case, if you go to either the Amazon show or the Amazon echo plus product page and then look for the link to “simple set up” devices, you’ll see a list of Zigbee devices including sensors and locks which are already a good 25% less expensive then the similar devices were last year.
Mass market zigbee is coming, I think, but via the stealth hubs and pretty simple rules options.
The Z wave alliance already said about two years ago that 2/3 of zwave device sales were for professionally installed systems like those from ADT pulse Or the various cable companies. The “do it for me” options. Again, much simpler rules choices than the standalone DIY hubs, but a larger market, In the millions of households.
So zigbee lives on, and Z wave lives on, but it may be that the multi radio DIY hubs are just too complex for the typical consumer.
I think any “hub” that is cloud based has no businesses in home automation.
From conversations with sources at each of the Big Three listed here, I believe they see value in local connectivity and acknowledge that 802.15.4 has better reliability, coverage, and lower power consumption than Bluetooth. Bluetooth 5’s standardized mesh is not reliable enough for door locks and security applications because it lacks network layer acknowledgements. It’s essentially “fire three time and forget.”
Today, Amazon’s strategy is clear with the Echo Plus. Google’s strategy is to integrate with the DIY hubs like SmartThings and Wink to enable door lock and security sensors (or guide customers to Nest’s very good line of first party hardware, which uses 802.15.4 heavily). Apple has dabbled with Bluetooth, but from conversations I’ve had, they recognize it’s a subpar experience for many users because AppleTVs and HomePods are often located in basements and media centers/closets far from door locks or the perimeter of the house where you’d want to install security sensors. That’s why they’re involved in Thread.
There are already triple protocol SoCs from Qualcomm and Unex that support 802.11, 802.15.1/Bluetooth, and 802.15.4, and I think these parts will start to appear in more Big Three hardware teardowns over the next few years. I’m aware of at least one other large silicon provider that has a triple protocol chip in development.
Z-Wave’s days are numbered, I’m afraid: single source silicon that already costs $1-2 (30-50%) more than an equivalent 802.15.4 part. None of the Big Three want to touch it in my conversations.
Who says that Samsung’s (ummm… “SmartThings”) strategy is a “traditional hub”? That officially changed just over a year ago at SDC 2017 in San Francisco. Samsung officially announced that they were consolidating their IoT initiatives (including all connected appliances, TVs, phones, etc. – all *hubless* by the way! – to use the “SmartThings Cloud” as the core of their platform. The Hub, in turn, would simply be a software concept. Just like Alexa is embedded in thousands of different pieces of hardware, so can a SmartThings Hub; including the Nvidia Shield, ADT Alarm Panel, Samsung’s mesh WiFi system, and more to come.
So you’re right: Smart homes can exist without dedicated hubs using efficient low-power, small packet, lightweight, self-meshing radio protocols; but such hubs are still entirely appropriate (and affordable – $70 for ST Hub 2018) to be added when consumers are ready to take their piecemeal smart homes to the next level. And if it turns out that WiFi, Bluetooth, powerline, or some other protocol supplants ZigBee and Z-Wave, the SmartThings Cloud will still be the *foundation* of Samsung’s IoT and consumer smart home platform.
I think the death of controllers is overstated.
The death of the first cloud controller may wind up saving the controller market in general. Lots of consumers will be ticked and not keen on moving to another cloud-based solution. Yes, this happened to Revolve, but Revolv was not a big player at the time. The impact of smartthings or wink shutting down would be seismic compared to Revolv.
Google is a serial project killer. They will manage to create some HA widget with decent market penetration and then kill it eventually.
I also expect the desire for a local controller to rebound after we have the eventual network outage that cripples a significant number of these cloud-based solutions nationwide. It might be AWS going down, DNS cache poisoning, DDOS or the classic “moron with backhoe” severing a major network interconnect.
The other eventuality is that people get vendor lock in and the vendor suddenly charges them money. These clouds are not free and eventually Samsung has to make money on smartthings.
At that point, some percentage of people will decide it’s worth it to buy something that will survive a change in corporate goals.
It is inevitable that controllers will be an ever shrinking percentage of the market, but I expect that will be due to market growth rather than declining sales. In other words, I expect dedicated controllers to be a niche product in very large market, much as home-assembled PCs are a niche in a billion-unit market. And yet Intel and AMD still put out enthusiast motherboards.
I used Wink in the early days and then used smartthings. I moved my Lutron and Hue lighting to Homekit for a while and then was enticed by the Google Smart Displays/Home Hubs. I moved the family to Android phones (Pixels) and sold my Apple speakers and bought some of the Google displays. I got a good deal on the ADT Smartthings system in November pre Black Friday 2018 but it turned out to be a disaster when I setup alarm monitoring so I returned it and went back to a Wink 2 hub. It has been rock solid and I decided to go with Nest for security and cameras and the two work really well together. I am really pulling for Wink and hope they stick around.
I started my journey, young, naive, and full of hope and wonder with a Samsung Smartthings v3 Hub about 6 months ago. I chose that particular model for a very specific reason, I fancy myself a “Samsung Loyalist”. This is not blind obedience like many of our distant, yet just as passionate cousins, the “Apple Loyalists” can be. Samsung has demonstrated a true willingness to listen to customers, what they want, and deliver a solid durable product almost everyone, albeit not without some missteps (Galaxy S6 series of phones). I knew the smart home hub market had finally matured to a point where it could enable me to live, at least to a degree, the future I used to daydream about. Naturally I didn’t think twice before I put my trust into a company I knew would deliver.
And deliver it did, even I had absolutely no issues pairing the first two Sylvania Outlet Smart Switches to the lamps in my livingroom. I was so impressed in fact, it would be weeks before what is apparently a universally experienced emotional tug of war in one’s mind everyone must reconcile with would creep quietly into my thoughts. I’m talking, of course, about that decision which is nothing less than the point in which a Smart Home toddler becomes the Smart Home Badass we all strive to be: Which is the better option, smart switches, or smart bulbs!?? I chose the oft considered option C, and use what makes sense.
I do lean to the switch side though, for no other reason than with all things being equal, I will always go with the more permanent option, or at least what I perceive as permanent. For instance, I’ll always go for a GE Z-wave wall mounted toggle switch with a dumb bulb, and while there is a story of failure with those, me, and a three way switch circuit installed by an apparent color blind electrician, we can save that for later.
Let me just say quickly, I know that GE, Honeywell, and a few others all source the switches from a specific manufacturer, whose name escapes me at this point. But if you’re reading this, “THANK YOU” for designing it with common sense in mind. In contrast I had 3 switches wired with the old X10 switches and paired with my garage door opener. I’m sorry, but I do not accept the “it was the first try” excuse to justify what was one of the most unintuitive designs I could ever actually think of. If a non-savvy guest needs to ask how to activate a Smart Switch in my house, the failure is self evident.
I’ve the first 3-4 months, I started building, one device at a time, what I feel is a respectable level of automation, not without some frustrations, but vastly greater levels of success. I would create a scenerio in my head, always tag on a random device to any and all Amazon.com orders, and make it work. My Smart Home personality traits started to slowly develop, finding that is partial to Z-Wave over Zigbee, I chose Google Assistant and silenced her arch rival, Alexa who snuck in with an Ecobee 4 thermostat, and found what worked and what didn’t.
Like I said before, there are just certain key tidbits which for one reason or another are not registered as demanding my attention. The most impactful of this journey, which also sounded the first death knell of my SmartThings Hub, was without a doubt very disappointing, but provided a very astute lesson in what those terms I had glazed over in the forums up to that point: Local vs. Cloud-based processing. Up to this point, I never really cared, I push a virtual button on my cell, and lights turned on and off. Any lag went unnoticed by me, everything worked, so whatever.
That was all nice and dandy, until I got a LITTLE to ambitious with my ideas. I live in an area where the Internet service probably drops off more often than Somalia. In most cases, simply recycling either the Cable Modem or my Wifi Router fixes the issue, but at that time, that would require my waking over to the shelf, and doing so manually with my hands, and as you’re all now thinking to yourselves “that just won’t work for me”.
You probably can see where this doomed road leads, but in case you are still curious, allow me to wrap it up. My very clever idea, which I was quite proud of, was to find a well rated smart power strip and assign every plug to a specific component, and then operate them independently from the others. I could cycle my modem’s power while in the comfort of my chair 3 feet away. There was an actually less-lazy concern, though. The internet didn’t seem to care whether or not I was home when it would decide to hiccup. My intent was to automate the entire process so I could ensure that my Wyze came were not disconnected and provide me the opportunity to look at my dog.
I ordered the power strip, and in 2 days, the nice Amazon people had me a brand new Meross WiFi 5 plug power strip, WITH 2 USB Ports to bat! I was so happy and proud, literally up to the point where I manually triggered the sequence I had been working to refine in the Tasker App for the previous few days. See, I hadn’t at the time given any thought to what would trigger the automation which needed to be based not on Wifi signal loss, but instead on network continuity loss. A task easier said that done. But that wasn’t what cloud me in to my error. The very sharp audible “CLICK” which had just indicated to me that the sequence had successfully begun……. Had begun…. Begun.. And it was then that I started to comprehend what Local vs. Cloud meant. It meant that when my SmartThings Hub took the first step of turning off the Modem, it could not know when to turn back on because that signal from the cloud was unable to reach its intended recipient.
Over the past following 3 months, I began to understand the limitations inherent to devices which rely completely on network connectivity. It didn’t take too long before I decided that brand loyalty be damned, having even a reduced level of automation quickly became something which I chose not to accept. Firth the exception of one Driveway alarm project, I have every one of my devices now successfully migrated over to my new Hubitat Elevation Hub, which boasts a local based system relying largely on Z-wave & Zigbee.
There are many things I miss about Smartthings, not to mention the fact that while this migration was being executed, Samsung released a new and unnoticed SmartThings app with since Gucci-cool features. This is where Hubitat really needs to work and improve moving forward. The UI is very clunky and dated, not to mention a very unintuative rules developer which has driven me to rage on many occasions. I still have automations that will not, for some stupid reason, do what I want,despite my attempts to shame it by reminding it that Smartthings never had a problem doing it!
So bottom line, decide what is important to you, and what you can compromise on.
Smartthings could do the job you wanted (maybe a better modem could have too), but you pretty much need, well, smarter things. A more programmable local controller can put the core intelligence in the controller. But you really needed a smart power strip that coukd be sent a single command to sequence power.
The demise of the hub is due to software. I have a Google Home Hub and all it will do is what you tell it to do. If you have any sensors, such as motion, water, door, etc. it has no means to take an event and do anything.
I have also tried Hubitat and Samsung Smartthings. I wanted to sense motion, turn a light on for 5 minutes, and turn it off. In addition, I wanted to control devices only Google was compatible with and speak back through Google from Samsung (send it text to speak). This should be a very simple workflow. If there is motion, turn on the light for 5 minutes, then turn it off. Nope, you have to learn groovy to do this or use IFTTT, and I am not sure IFTTT will even do it.
I started learning Groovy, I tried setting up Google Relay (Node JS server, Webcore, install scripts with massive errors).
I do not believe the average user is going to do all of the above to get the Smart home they want. You should not have to be a tinkerer to figure this stuff out, nor should you have to spend endless time on forums where instructions are incomplete at best.
That, I believe, is the demise of the hub and something very easy they could fix.
Funny thing is that the simple application you had is well served by a Wink “robot”. Winkmcan act on all sorts of triggers, and react including time-based reactions. It’s time isn’t bad… Turing my outdoor lights on 15 minutes after sundown is based around its built-in tracking of sundown and sunrise. It can’t do complex things, but they do have basic automation that’s very easy to use. But so slow to react.
No need for Groovy. Webcore (designed for Smartthings) is much simpler, and can do pretty much anything you want. The new automation engine for the Smarthings app is getting more sophisticated (the developer of webcore now works on Smartthings), but it will be a while before it can touch what webcore does. For instance, I wrote some fairly simple webcore “pistons” which monitor my sump pump and washing machine (separate pistons for each) and lets me know if the sump pump did not run on a day it was raining and when the washing machine starts and finishes (and even if the door has been opened since finishing).
Missing from the discussion about hub or not to hub are the fascinating opportunities and potential to be found by exploring one or more of the existing open source home automation projects. While these are perhaps not yet ready for general use, they are being used by thousands of both technical and non-technical “DIYers”. Users’ feed-back plus the open software development contributions of many of them around the globe are moving these projects rapidly forward. An on-line search on, for example, “open source home automation” will deliver useful links.
Among the types of benefits such projects deliver are broad-based and long-term support independent of any commercial company’s decision process, ability to integrate with many devices , and each user’s right to decide which use cases to implement and what decision rules to apply.
Fascinating and long read covering 2 years of evolution. Sometimes it might be important to get e fresh outside perspective. Enter Europe. It is somewhat disconnected from US thinking in many ways leading to different emerging solutions: Homey, Conrad Connect, Homematic. Homey does all what the original article suggest in a pragmatic approach. I hope they enter the US market to break the fiefdoms of AWS, Goggle and Apple. Such is possible, see Clayton Christensen’s Innovators Dilemma, the emergence of Zoom over Cisco and Polycom etc.
Google’s products and home hubs do have low power radios. They are “Thread” radios.