First and 2nd gen Nest Thermostats will lose support in Oct 2025
arstechnica.com
arstechnica.com
Firstly "lose" is passive. Google is ending support.
But most importantly "support" sounds like they just won't be fixing bugs and security problems. What is really happening is that Google is disabling all online functionality.
"Google is shutting down all online functionality for 1st and 2nd gen Nest thermostats in Oct 2025" would be a much clearer headline.
I'm guessing since the generations being shut down are explicit, I'm assuming there are more generations that aren't being shut down, so the simple answer is that they want to sell more of those devices, basically force people to upgrade.
So indirectly, letting people still use the devices locally, would mean they'll sell less of the newer devices, which probably have better margins, slurp up more data and will eventually be shut down too.
Indeed, had they started to support a product they would never say “15th gen Nest thermostats found support”. You know, they were walking through the woods and stumbled on some Google support. Isn’t that a nice coincidence.
I'm wondering if an approach to saving older Nest thermostats and reducing ewaste might be:
- Using the older ‘unofficial API’ (from around 2014) assuming it still works, and integrate it with Home Assistant (see GitHub - gboudreau/nest-api: Unofficial Nest Learning Thermostat API ( https://github.com/gboudreau/nest-api ) & GitHub - wiredprairie/unofficial_nodejs_nest ( https://github.com/wiredprairie/unofficial_nodejs_nest#readm... ) )
- Encouraging folks to put older and unsupported Nest devices on a separate wireless network that doesn’t have internet access but access to a Home Assistant server for remote monitoring & control.
- Facilitate a way for Home Assistant based HVAC control setups to work with energy company savings programs where appropriate (in areas of the US this is fairly common, but they may only be supporting major brand vendors so far)
The solution is to stop installing advanced touchscreen computer systems where they don't really belong.
I've never seen a Honeywell thermostat fail, get hacked, or "lose support". I dont have to worry about what will happen if there is a weird software update or power outage. There are physical switches that configure the current mode of the system. Only the temperature setting is volatile, but it defaults to reasonable settings depending on the current mode (60F for heat, 80F for cool). Because of this simple design, there is no option to install a battery. There's nothing to remember.
The thermostat should be a remote display and a sensor. The real brains should happen on a controller near the heating system, that can be swapped in and out as time goes on.
Newer A/C units have EC motors with variable speed motor controllers and newer furnaces will have an EC motor blower fan and lots of factory control points wired up to a controller, mostly for the burner (and related gas/burner safeties).
Still, even in newer furnaces and a/c units, the thermostat is the main controller as like you said, the calls for heat/cool are the only input to the system.
Commercial HVAC controls for air handlers, boilers, and chillers have a lot more control points, an AHU can have 40 or more control inputs and outputs tied to it. Discharge and return air temp, multiple differential pressure sensors, static pressure, outdoor air temp, outdoor air pressure, duct avg temp, etc. These systems are much more complex than any home HVAC setup and usually have multiple controllers networked back to a master controller which orchestrates everything.
What is the benefit you see of moving the swappable brains closer to the heating system? Doesn't that actually make the whole system more complex because now you have two separate devices where there is currently one?
They will still work for turning the dial locally to control the temp. They just won't be IoT connected.
The 3rd gen Nest thermostat was released in 2015. The 2nd gen and 1st gen were released in 2012 and 2011 respectively. A decade lifespan for a device that gives me metrics on my climate control and lets me tune it from my phone is pretty solid in my book.
That's the attitude I take issue with. The device should work as it did at the time of purchase in perpetuity. If it can't, because of the architecture of the product (i.e. needs third party servers on the Internet) it shouldn't be marketed as being sold, but rather rented.
Acceptance that products we buy will just stop working (or, in the case of this product, losing significant functionality) is only encouraging companies to ratchet down lifetimes and hurt consumers more.
If all the shelves at Lowes that sell network-connected IoT thermostats said "10 year rental price" instead of "purchase price", the average non-nerd consumer wouldn't change their shopping behavior. The nerds... maybe they would? but they're the ones who already realize the impact of a device needing cloud connectivity to function.
The reality is that vendors cannot and will not offer "in perpetuity" commitments to run the cloud portion of their gear. For strategic reasons like "they want to deprecate legacy endpoints and functionality on their end" but also unintentional reasons like "the company went out of business".
Everything you buy has a lifespan. Physical objects degrade over time, objects with cloud connectivity rely on external functionality, etc. Picking between them is part of the decision making process: do you buy the cheap furniture that falls apart the first time you move? Or the more expensive furniture that lasts decades?
Should smoke detectors be listed as "rentals" because they degrade and have to be replaced?
Exactly. Consumers have no way to know, right now, if something they "buy" is truly a purchase or a rental. They can't make informed decisions without that knowledge. A technical person can tell the difference, but the average consumer cannot. That's why the distinction needs to be explicit.
I buy a smoke alarm with a fixed 10 year life to avoid battery replacement. Others buy models with replaceable batteries, still others install wired alarms.
You should disclose the lifecycle at sale so consumers can make a decision. I’ve been in the computing infrastructure business for many years - I can model out the expected life and cost structure for all of my gear.
I’m not sure that an average consumer is even aware that some smoke alarm sensors degrade with time, or not trigger at all types of fires, there is probably some logic in forced unit replacement every decade.
It may be an example of an unintended consequence - to be fair to the Nest people, when you’re releasing a new product, you may not be thinking of end of life. Today we should be stamping a support date on them - Google is pretty good about this with Chromebooks, for example.
Right. That is precisely why nothing that is meant to last should have any kind of dependency on "cloud". It must run locally only, so there is nothing that can be deprecated or dropped.
My mid 90s thermostat will continue to work independently of whether the original company that made it is out of business. Everything needs to be like that.
Unless you're living off-grid, I think there's still some benefit to having one of the largest energy loads (building HVAC) coordinate a bit with utilities to ensure there aren't blackouts and such. Even better if there's an approach that works with open-source solutions like Home Assistant.
Where I live staring at the grid load and available energy is a thing during heatwaves and other inclement weather, and the utility generally issues "energy rush hour" limits to smart thermostats to help ensure grid reliability. The utility also incentivizes this I think by providing a small discount on the utility bill.
Target a Specific Ask: Focus on allowing basic remote control via a local or self-hosted solution, not full open-sourcing. Examples: Release a local API with secure authentication Release a Dockerized server image for basic relay control Extend cloud access via an open licensing model Find Allies: Partner with groups like iFixit, the Repair Association, or the Electronic Frontier Foundation (EFF). Reach out to smart home and privacy-focused influencers or YouTubers. Launch the Petition on the Right Platform: Change.org, iPetitions, or Action Network are common options. Consider also running a campaign through Reddit, Twitter/X, or YouTube. Provide Testimonials and Data: Show how many devices are still in use. Document how much e-waste this move could generate. Share personal stories of user frustration
Nothing will compare to ads in terms of margins. It makes no sense for Google to work on anything else, because eventually some VP will see compare the margins across business lines, and the decision is obvious. Chop everything that does not sell ads.
But if you buy a Nest, you're buying a network-connected device. A decade replacement cycle for network-connected computing devices is not crazy.
It should be, though. Okay, I get spare parts not being available after 10 years, particularly the "smart" chips simply don't get produced any more. But firmware and API support? How fucking hard can it be for a company raking in 90 fucking billion dollars in a single quarter of a year to keep a dozen people on payroll to keep the old servers, APIs and build tooling for the firmware up to date?!
[1] https://finance.yahoo.com/news/google-stock-rises-after-it-b...
https://hackaday.com/2019/02/27/hack-my-house-raspberry-pi-a...
The crazy part is that consumers are allowing the idiotic product lifecycles from the IT industry destroy products that, otherwise, had long and predictable lifecycles. It's shocking how the IT industry is successfully conditioning people to think it's normal to throw away otherwise working gear because "upgrades". The industry should be held to account before they take any more consumer goods hostage.
It feels like your actual beef here is that the majority of consumers are interested in buying cloud-connected gear that degrades when the vendor moves on, but you're framing their choice as "crazy".
My concern is that consumers don't know what they're buying. They don't know what "cloud connected" means. I don't care if people but this stuff, I care that they know what the bargain is that they're making.
How old is your toaster?
What's your reasoning behind that belief? A device being "modern" doesn't automatically mean it has to be user hostile. That is a choice.
Sure it is, and it’s why we desperately need consumer protection laws mandating lifetime security updates for IOT devices.
My grandparents had a 100 year old home - some of their light switches were well over 30 years old. Why should I be forced to replace a smart light switch before its useful life is over? Because some company wanted built in obsolescence so they could sell more product?
Edit: and to be clear, I’m potentially fine if the vendor takes the approach of open sourcing all firmware as a way to allow lifetime updates if they need to get out of the product for some reason and can’t maintain it themselves. “Reasonable accommodation”?
To maintain security and guarantee continuing safety, CarOS™ will now drive you to your local showroom where you can purchase updated hardware.
The entire world offerings are moving to this type of model under the part guise part reality of tech dependencies that just don't exist in a stand alone product.
Stand alone products are disappearing more and more every year that passes. It is certainly no accident.
- could no longer do code deploys on internal WebApps that were on old Python 2/3 images. Had to stick with that revision until we migrated to newer versions.
- had to migrate to a newer firebase messaging API recently - search features mostly killed on GCP, which is ironic. I guess we're supposed to build our own search or use a 3rd party? Postgres/Spanner full text search it is then.
Looking forward to having to migrate my gen1 cloud functions too! Keeps me employed I suppose.
Because the money is made by selling the products, not by supporting the users.
Fundamentally the idea end point devices should be connected to the Internet is wrong, and this means the idea the manufacturer should run any associated cloud service or be responsible for "security updates" is also wrong.
The only long term solution to this is that "smart devices" not be given direct access to the Internet, where the security problems are, and live instead in their own strictly segregated networks controlled by something conceptually like Home Assistant but where it also acts as a broker between all local devices, with a whitelist only policy for networking access. Until that becomes the default mode for deploying these things the consumer IoT space will continue to fail to meet expectations.
Then, years later there was the push to move people to the Google Home app from the Nest app. I still have both. But Nest was better before Google bought them. I had better customer service when I called them for help wiring my thermostat.
This configuration has allowed me to control the single-zone furnace using any or a combination of 'virtual' thermometers to turn the relay board on/off depending on conditions. I left the real old honeywell thermostat on the wall as a fail-safe situation in case the system comes crashing down, to prevent the house from freezing up in the winter -- Just set at a low minimum.
The downside of this is I either have to directly adjust the temp via phone interface / tablet interface / or google-assistant interface, but when that is done, it is all done locally. Simply toggling a simple relay board that just connects the 2-wire thermostat feed like a typical thermostat does. It has worked great!
In addition, one of the heat pump units (MrCool!) in the house was able to use a cloud-free dongle to bring it's controls into local-only control as well via home assistant, as well. This can also be controlled via phone / tablet / google home interface (I subscribe, I feel home assistant is worth supporting).
All of the local only devices are able to be accessed from afar using tailscale into the local network, with subnet routing.
When you have something like home assistant controlling it all, it's quite easy to set up situations like 'when the outside temp is less than 15(f)', turn the heat-pump off, and use the oil furnace instead, etc.
It seems the commercial control boards that do things of this nature cost a heck of a lot more, and effectively do very similar things.
It's a good time to start really considering a more local-based solution if you have the time to get the initial configurations all sorted!
[1]https://news.hydroquebec.com/en/press-releases/2172/hydro-qu...
The whole idea is they can turn down your thermostat and sell the electricity to the US instead, or it was. Somehow they also encourage the adoption of electric cars, while also claiming we should save electricity. It's a very confused message.
Its heating curve is corrected by a PID. Room temperature precision is within 0.1 Celsius.
Less gas consumption, lower bills and higher comfort :)
Wait. Oct? They run Windows 10?