Today, we’re excited to announce early access to a new version of Chrome OS bringing the benefits of Chrome OS to PCs and Macs. Chrome OS Flex is the cloud-first, fast, easy-to manage, and secure operating system for PCs and Macs. Learn more below, try it out, and share your feedback to help us shape this product.
It’s basically Chrome OS for any PC or Mac you might have lying around. Other parties offered similar Chrome OS-based systems for non-Chromebook hardware, but this is the first time Google itself offers it as a first-class citizen of the Chrome OS ecosystem.
I will be honest. I was worried when CloudReady was initially acquired by Google. Fortunately it did not go the way of CentOS/RedHat, but is becoming even more functional.
For those who are curious, the USB can be used as a “live image”, and will allow logging into the Chrome OS (Flex) without modifying the hardware. However I think it does not store the accounts, hence need to do 2FA after every boot.
Yeahhh…here’s the thing, its Google, a company notorious for pulling the plug on projects that don’t show a ROI pretty sharpish or in this case since its pretty obvious from the website they are targeting enterprise I smell a coming paywall for anything more than the most basic of experience.
And with it running on bog standard hardware as opposed to the more locked down ChromeOS hardware why would you just not use Linux since that is all ChromeOS is really? I mean its not like there isn’t user friendly super low resource Linux distros, I currently have Zorin running on my 11 year old AMD based EEE PC netbook and that thing wasn’t a powerhouse when it was new but the 8GB of RAM and $20 SSD I slapped in trying to get Windows to keep going (You can go get a sammich and eat it and Win 10 won’t have finished loading the desktop on boot) are frankly barely getting used and the whole experience is actually quite snappy.
So I really don’t get who this is aimed at really, having worked in corporate I can tell ya nobody is keeping hardware past the lease, at least not here in the states as you can no longer take depreciation off your taxes and its a lot bigger PITA to actually have an IT dept fixing things in house than just pack it off to Dell or HP or whomever, and while having done a bit of work at schools they can hang onto gear longer frankly other than the teachers nearly all of them have gone Chromebook thanks to the pandemic and I kinda doubt this is gonna be a backdoor way to get more life out of aging Chromebooks LOL.
For now i will stay with the “legacy applications” option and will leave the “cloud first” option to the hippies.
Which brings the question: Since Chrome OS runs Desktop Linux apps, can you run Wine with DXVK as a Desktop Linux app on Chrome OS?
Imagine if the Year Of The Linux Desktop happens thanks to Chrome OS and SteamOS… I mean, why not? A Year Of The Linux Desktop powered by Chrome OS or SteamOS is something I can reasonably get behind (I say this as a person who has been unconvinced by Desktop Linux for the past 20 years or so). I never really “got” the purpose of companies like Canonical who don’t really have the means to produce a desktop OS suitable for public consumption and are mostly just repackaging stuff from upstream. Companies like Canonical are basically like: “Debian is perfect, it just needs some lipstick, and we are the ones to provide it”. I mean, no, that’s not how it works, look how much work Google and Valve are putting into Chrome OS and SteamOS to make them fit for public consumption. Of course, this means Google and Valve bake revenue-generating mechanisms in their OSes, which irks some linuxeros, but don’t worry linuxeros, a rising tide will benefit all boats.
Yes, Wine and Steam already runs on regular Chrome OS (but did not check this new one).
There is even some hardware acceleration support, including vulkan:
https://chromeunboxed.com/how-to-enable-vulkan-crostini/
But, the default Linux container has a small disk limit, however it can be expanded.
Canonical did some good work in the beginning. They added a lot of polish to the Linux desktop. They did the 10% work no one was doing at the time. Basically, they filled the niche Google and Steam are now. However, they didn’t have a product which could subsidies the desktop stuff the way Google and Steam do, any sort of great vision, or any sort of idea about what enterprises need in a desktop OS. (Hint: It’s all about fleet management, and RH is the only one of the old guard who kind of has an idea about how to sell Linux desktops to large businesses. Google has figured out fleet management, and that’s the big selling point of ChromeOS.)
This is why I never “got” companies like Canonical: The Linux Desktop needs (needed?) much more than some polish (lipstick): I am talking about things like improvements in the graphics stack, audio stack (so HDMI works out of the box and pulseaudio doesn’t mess up volume levels), app distribution, win32/win64 compatibility, and driver support (PROTIP: bellowing at companies to write good drivers isn’t a plan, offering them massive chip orders like Valve did for AMD is however a plan). None of those “distribution” companies had the means to make any of that happen, much less all of them.
This is why I never got the business plan of “distribution “companies (Canonical, Mandriva, Linspire etc): Their business plan was basically like: “We don’t have the money to fix what really needs fixing in Debian/Fedora, much less improve win32/win64 compatibility, and driver support, but if we put some lipstick on Fedora/Debian, we can somehow make a sellable OS out of them”. I mean, no. You have an OS that’s less unsellable than Debian/Fedora, but still not sellable.
Another thing I never understood is what those companies pretend you can just drop a Live CD on a Windows PC and it will work (which is a pipedream considering how diverse the Windows PC ecosystem is), instead of having “official hardware” that’s guaranteed to work (like Chrome OS and SteamOS do) and classify everything else as beta, but that’s another rant for another day.
kurkosdr,
I appreciate where you are coming from. But you’d be surprised how useful those linux Live CDs were to me even as a windows administrator. I would use a linux live cd to download network drivers for windows, haha.
kurkosdr,
Canonical’s free Ubuntu CDs were how we introduced Linux to many of our peers and students. I used to order a pack every year, and give them away.
It got so much attention, Microsoft learned about it, and gave us free Windows ISOs (but not physical CDs) to distribute to students.
(Now I think it is free to any @edu account, but not sure).
You say you find Ubuntu to be just “lipstick” and don’t really get it on what it is all about. On the other hand you would get behind Chrome OS Flex. You do understand that Ubuntu and Chrome OS Flex will more or less have the same issues on your hardware if there will be any? Regarding device drivers support. Hence if one is just lipstick, then they both are just lipstick. But obviously neither of them is just lipstick.
It’s really quite usable. It’s the best it’s been in the 20yrs I’ve been using it. There is 10% stuff which needs to be taken care of, but it’s the best it’s been in 20yrs.
Are you using Nvidia hardware? I’ve kind of taken for granted these are going to work for a few years now with Fedora.
Flatpak is probably the App Distribution method of the future.
Win API compatibility is a problem handled by Crossover via Wine, and that has nothing to do with Linux. Wine shouldn’t exist, but that’s not a popular opinion.
Driver support could always be improved.
It’s the other way around. AMD already had FOSS Linux drivers, and Valve picked AMD because they can work on the driver.
That had nothing to do with Valve. If I had to pick something, it was probably the lucrative workstation and server market Nvidia has locked up.
Desktop Linux is pretty much RH and Suse. Those two have done the major work Valve and Google are benefiting from. Both put major resources into desktop Linux moved some big mountains.
Canonical did start the ball rolling and showed it was a viable idea with a little work put into it.
There were many problems back when Desktop Linux was a hot idea, and most of the problems weren’t solved until Electron showed.
That’s how a startup works. Start small and eventually get money to fix the problems. System76 for years installed stock Ubuntu on their rebranded Clevo stuff before designing their own desktops and customizing Ubuntu. They still don’t contribute nearly as much as other players do.
I complain about hardware support, but it kind of just works most of the time. Especially if potential lawyer traps are ignored. Aside from Apple, most x86 OEMs aren’t that creative.
I remember running Knoppix on my Windows desktop circa 2003. 🙂
I use Linux on desktop and desktop applications that are installed on my computer. Don’t plan to change that in foreseeable future. If other people want to be part of some “cloud first” walled garden. By all means feel free to do that. As it’s not like majority of people hasn’t done that in the past few decades anyway.
I thought Google was offering ChromeOS based cloud desktops, and was kind of interested, but that’s not what this is.
I’m going to say, “Nah. I’m good.” Of course, give me six months when Mozilla announces they’re going to switch to Chromium.
This would be slightly more interesting if they had full support for Intel based Apple hardware. I have a 2014 Mac Mini which runs Fedora pretty well, but the wl drivers for the Broadcom WNIC are craptastic.
Hmm. Some random thoughts and questions:
Will this have any advantage over traditional Linux distros in supporting problematic WiFI chipsets like Broadcom and Realtek?
Will there be a simple non-geeky way to install Linux desktop apps?
This would have massive appeal if they offered Android app and Play Store support, but it doesn’t. Seems like a huge missed opportunity.
And the big question: How long before Google with its attention span of a 2 week old puppy loses interest in this?
I just tried it live, without installation.
1. Problematic wifi chipsets are still problematic. My desktop’s cheap usb wifi adapter was not recognized. So I was unable to use the os on my desktop at all. My old acer laptop’s wifi was recognized, but its usb 2 ports weren’t. I was only able to boot using the single usb3 port. So, hardware support is still hit-and-miss.
2. “Non-geeky” and “linux” do not belong to the same sentence. Not in this context… not in any other…
3. There’s chrome web store, which contains some apps such as a VLC player with 2 star review rating. But yeah, the lack of Play Store is a huge deficiency. On the other hand, this OS looks very promising for use as a kids’ computer. You can set up a child account, just like you do on Android. It will give you a more or less safe environment for the kids to use office apps, basic browsing and youtube. I’m looking for some dirt old netbooks with 4gb ram, to build gifts for my nieces.
4. And the elephant in the room… Google’s attention span… Well… Hopefully this project will not be abandoned before the glitches are ironed out. It is obviously a work in progress, especially on the hardware support front. But realistically speaking I don’t expect Google to come up with drivers for 10 years old Realtek chipsets.
2: ChromeOS is locked down and pretty nerfed. It’s more of an appliance distro then a traditional Linux distro.
3: The Chromium project indicates Play Store and Android app compatibility, for certain devices, has been a thing since 2016. https://www.chromium.org/chromium-os/chrome-os-systems-supporting-android-apps/
4: Google is making money off of ChromeOS’s Google Workspace integration and Chrome being surveillance tech, so it should be okay.
It may not be Linux based in the future, but ChromeOS will probably exist.
cevvalkoala,
This is what I expected honestly. My experience with chrome OS is that it has less peripheral support than linux, although obviously if you stay within the chromebook’s officially supported hardware ecosystem that shouldn’t be a problem.
It depends. “Linux” is rarely the problem, it makes a difference how it’s bundled up for consumers though. Ordinary users frequently use linux without any knowledge that they are doing so as with android, chromeos, steam os, roku, etc. Even a linux computer isn’t that difficult to use on a supported system. The vast majority of linux compatibility complaints would go away if you could limit users to supported hardware combinations. The problem of course is that supported hardware is *not* the way people get involved with linux computing. Instead users generally install some distro on a random windows PC that has never provided any linux support. This is all too common and the result is users becoming alpha testers for their unique hardware configurations. It’s no wonder things can go wrong 🙁
To my knowledge google doesn’t write any of the drivers. They’re using the drivers in linux itself. They basically choose which drivers/modules to include in their linux build.
Unknown. Its hardware support is the same as Linux, for better or worse. This might change in the future, but it’s unknown at this time.
No. Apps have to come through Google, and the Linux desktop applications we know and love aren’t really a thing. ChromeOS is very locked down and nerfed. People speculate about Apple locking down macOS, but ChromeOS has been that way from day one.
People have jailbroken ChromeOS to get regular Linux distro functionality, but being a regular Linux distro isn’t what ChromeOS is about.
Flatland_Spider,
That restriction is no longer true. Chrome has been able to run native Linux applications for a while. For example, here are instructions for Visual Studio Code:
https://code.visualstudio.com/blogs/2020/12/03/chromebook-get-started
And for those who wonder “how the sausage is made”: https://chromium.googlesource.com/chromiumos/docs/+/HEAD/containers_and_vms.md
Per the VSCode docs Crostini is needed.
Per Google Crostini is for devs.
Flatland_Spider,
I’m glad it’s possible, but sideloading on chromeos still isn’t as friendly as on android (or a desktop distro for that matter).
Am I correctly recalling that chromebook users had to wipe their devices clean deleting everything in order to enable this? If so IMHO it’s still kind of a high barrier for average users who may not even know about this feature until it’s too late and they give up with sideloading linux software because they don’t want to do the format & reinstall dance.
Flatland_Spider,
I think you are confusing Crostini vs Crouton.
Crouton is a “dev mode” distribution that has root access to bare metal.
Crostini is a container based distribution that runs in a VM, but runs at near native speed with partial USB and GPU acceleration support.
It runs on normal mode, heck, I am writing this inside a Chrome browser running in Crostini on Chrome OS 🙂
@sukru
You are correct. I am mixing the two up. 🙂
That’s cool. I was thinking it might be something like the toolbox utility or a systemd nspawn container. A little container designed for interactive work on top of an immutable host.
https://github.com/containers/toolbox
@Alfman
I was mistaken. 🙂 I was thinking of Crouton.
I’m not sure. I can’t find anything about that. Maybe because developer mode needs to be activated?
Anyway, there are probably better ways to go about getting Linux software on ChromeOS if they thought about it.
Flatland_Spider,
Yes, it is easy to get the two mixed, especially when it was not promoted well.
I don’t know how much of their work is generic. But it might be a good starting point for a general purpose desktop OS. (They already support apps integrating with the outer OS, so VS Code gets a separate window, launcher icon, etc).
As you mentioned “immutable host” + containers might be the future. (I still am waiting for Windows to do this on their end as well).
Sukru,
Fedora Silverblue is close to being an immutable host. The base is managed by ostree, and the applications are delivered via flatpak. Development work is done in mutable containers created via podman or toolbx. It’s been the closest FOSS OS to ChromeOS I can think of.
I need to do some more work to get comfortable adding things like codecs and drivers to the base, but Fedora Silverblue is interesting.
Flatland_Spider,
Interesting,
I had lost hope on CoreOS, but they seem to be using that for Silver Blue (for side note: CoreOS was a clone of Chromium OS base for containerized read only clusters. Bought out RedHat a while ago).
(Found this:
https://www.reddit.com/r/Fedora/comments/chhbeu/what_is_the_difference_between_fedora_coreos_and/ )
My only concern is (apart from project being very fresh), they don’t seem to support multiple copies of the apps/flatpacks:
https://www.reddit.com/r/flatpak/comments/rxna9y/running_multiple_instances_of_the_same_app_with/
I have duplicates of browsers for isolation. They seem to be lacking that feature.
Sukru
They’re related. 🙂 Silverblue is the workstation version of CoreOS.
I forgot about CoreOS being a fork of ChromiumOS.
True, there are still features being added to bring it up to par with the regular Fedora WS, but I’m assuming Silverblue is the future of Fedora WS, in some form or fashion.
I haven’t tried multiple application versions from flatpak. It’s not something which has come up, but that’s probably not a scenario which was envisioned. Or the answer is to setup a container with the software.
Hmm. Some random thoughts and questions:
Will this have any advantage over traditional Linux distros in supporting problematic WiFI chipsets like Broadcom and Realtek?
Will there be a simple non-geeky way to install Linux desktop apps?
This would have massive appeal if they offered Android app and Play Store support, but it doesn’t. Seems like a huge missed opportunity.
And the big question: How long before Google with its attention span of a 2 week old puppy loses interest in this?
That’s something I never understood: Why is this a requirement in the first place? Let’s be honest here, very few people will pave over the OS installation that their computer came with (it’s the reason Vista kept being a thing in the statcounter stats despite the fact there is literally no reason for a Vista computer to not be upgraded to 7). And if, as an OS vendor, you are selling Desktop Linux pre-installed, you can hand-pick the hardware to be Linux-compatible.
This “requirement” is something only a small percentage of users who happen to be nerds (and can pave over OS installations) care about, but it’s not a requirement for any OS distribution aiming for a wide market share. SteamOS is a good example: It doesn’t even pretend to be something you will pave over your WIndows installation with (not on random Windows PCs anyway), it comes pre-installed on official hardware.
But anyway, to answer your question, there is a hardware compatibility list: https://support.google.com/chromeosflex/answer/11513094
But I personally wouldn’t expect anything more than whatever the standard Linux kernel.
Buy your computers to be linux-compatible instead of demanding that linux supports every crappy network adapter out there.
Because people buy junk hardware then complain about it when it doesn’t work, and the OEMs put a lot of junk hardware out on the market.
Basically making the best of a bad situation.
This happens with official Chromebooks and Chromeboxes. This won’t happen with random hardware.
ChromeOS Flex advertises itself as an OS for general x86 hardware. The message is “Download ChromeOS Flex and install it on any random PC or Mac around!”, which means people are going to try to install it on all sorts of junk.
They also explicitly advertise Mac compatibility, and Apple used a specific Broadcom WMIC which doesn’t have drives.
Just one thing to keep in mind: People buy Windows PCs. Sure, the hardware is usually junk in a variety of ways, but my point still stands: People buy Windows PCs when they buy the aforementioned junk hardware. Paving over OS installations is a nerdy thing that most people rarely do.
Which is why ChromeOS Flex is a ChromeOS variant for nerds who know what they are doing, or for consulting companies that do the nerdy stuff on the customer’s behalf.
Still, I think Google is taking a risk by putting Chrome OS Flex out in the wild without the appropriate caveats. Their own compatibility list is indicative of this (see my previous comment).
True. People buy Windows more then they buy the PC brand.
Apple hardware is a special niche, and probably the better niche to target. At some point the hardware ages out, but would still usable with Linux. I”d like to snap up some cheap Mac Minis once they lose support from Apple. 🙂
I immediately though of some large business who wants to reuse their existing equipment and get the perks of ChromeOS integrating with Google Workspace. People can’t sign up for the demo without providing a company name.
Definitely. Releasing this to the general public is a bad idea.
Releasing ChromeOS compute instances hosted on GCP and accessed via RDP/VNC/SPICE/whatever is a good idea.
kurkosdr,
+1
I agree that would go a long way. I don’t know where the idea that linux needs to work with all hardware comes from, but it sets an unrealistic expectation. Look at apple by contrast, there’s tons of hardware they don’t support at all and yet most users aren’t complaining about it. They understand that they have to buy macos compatible hardware.
The idea that Linux needs to work with all hardware comes from the LiveCD concept. The LiveCD concept was, in turn, a concept born out of necessity: Companies such as Canonical, Mandriva etc didn’t have the capital to negotiate manufacturing deals (like Apple, Google and Valve do), and the state of their product (OS) wasn’t able to generate the required volumes anyway (see how few units the Dellbuntu netbooks sold), so they had to get their OS out there somehow.
Hence, the idea that you can just plop Linux on random Windows PCs, or try and convince companies like Dell to plop it themselves on random Windows PCs they make.
From a usability perspective, it’s a terrible idea, considering how diverse Windows PC hardware is. I believe Apple’s idea was the correct one, aka make your OS work well in a small number of devices and sell those devices in good-enough volume in retail channels. With Steam OS and Chrome OS, we finally got exactly that, that’s why I believe Desktop Linux finally has a fighting chance.
I am still not convinced about Chrome OS Flex though. It has the problems of the LiveCD concept (again, their own compatibility list is indicative of this). I believe Google should have made it clearer it’s for devs/nerds only.
This is BTW why I can’t understand how the idea that Linux needs to work with all hardware became a requirement. The LiveCD concept that gave birth to this requirement was a bad idea from the start and didn’t even put a dent in Microsoft and Apple’s market share.
This kind of stuff should have been marketed for nerds only, instead of being widely advertised. Personally, an official Ubuntu Live CD that I tried soured my opinion on Desktop Linux after it locked up my BIOS on my HP laptop every time I shut down (I had to pull battery and plug, until I figured out the first time I was pulling my hair) and I couldn’t get hardware acceleration to work on the laptop’s ATI Mobilty Radeon X1600 (a popular GPU at the time).
But you know, the Shuttledorks didn’t want to invest too much capital (for example on manufacturing deals) in order to have actual official hardware, so LiveCD it was.
Not to mention the real issue here, that the Shuttledorks didn’t want to invest capital in fixing the flaws of Desktop Linux itself.
But this is what happens when there is no consistent revenue source.
kurkosdr,
Personally I think that livecds worked pretty darn well all things considered. It was a convenient way to try linux without having to install it first. There’s something appealing about being able to try out a distro simply by booting up strait away without having to mess with around with partitions and install it before I can even give it a spin. Live CDs managed to get people to try linux who wouldn’t otherwise have an easy path for doing so. The one I used, Knoppix, was fairly well supported and had a decent user base.
So I guess I don’t have the same experience as you do with them.
Of course it’s helpful to target a small number of devices like apple, steam, or chrome os. But a big difference is that they’re selling the hardware. Linux as an entity doesn’t exist to sell hardware, it exists as an operating system that users can install.
The *ideal* scenario is that manufactures would support linux like they do for windows, but it hasn’t happened in decades and probably won’t happen. At least not before linux has critical mass…alas it’s a classic chicken an egg problem. Would you be a proponent of the linux foundation selling their own computers then? And let’s say they did, do you think people would actually buy their hardware enough to make a dent in the market? I don’t know, it’s an idea though.
It was Electron. That was the game changer. Also, maybe WebRTC and WebSockets.
The major problem has always been applications.
SteamOS uses Wine to work around Linux, and people are buying ChromeOS because of the management capabilities of Google Workspace.
Valve and Google are getting way too much credit here.
That wasn’t seen as a viable business plan at the time. Apple was a laughing stock for staying vertically integrated, and everyone felt MS was the better business model to emulate.
I think it’s from people using Windows as a measuring stick of Linux.
I would like Apple hardware to be fully supported by Linux.
Yeah, I don’t disagree with you. You’re kind of preaching to the choir here; I only run Linux and I’ve turned down some fantastic deals on otherwise excellent laptops but with crappy WiFI chipsets. I was just wondering if being backed by Google would mean that they invested any additional effort into negotiating licenses with the likes of Broadcom and/or fixing the garbage Realtek drivers and shipping them out-of-the-box. Pretty pathetic that they don’t; I distribute a somewhat popular set of spins built from openSUSE as a one-man project with an annual budget of $0, and I manage to include Broadcom WiFI support out-of-the-box thanks to a third party repo that packages it for the openSUSE kernel versions. In actual reality a lot of users already bought hardware with junky WiFI because they have no idea what is compatible with Linux or even the fact that different WiFI chipsets exist, and when they later want to use that hardware with Linux it’s always nice to support it if possible.
Is ChromeOS phoning home to Google? If yes: no go.
Wondercool,
I don’t like it either, but given that almost everything is phoning home these days I would assume the answer is yes.
I tried ChromeOS Flex on an unsupported, old laptop — a Lenovo Thinkpad X301. The laptop still works great but over the last few years Linux distros are getting a bit more robust than the X301 can handle so I figured why not try ChromeOS Flex. The installation process was challenging and buggy, especially at several points where text prompts weren’t visible unless you manually highlighted them with your mouse. However, once installed, everything seemed to run properly. I’m still testing it — haven’t looked into whether it can run Chrome’s Debian container, for instance. I wouldn’t however, as Thom writes, call it a “first-class citizen” of the ChromeOS ecosystem as Flex doesn’t offer support for Android apps (I’d be happy if I’m wrong on this).
Also, I’m wondering if someone couldn’t install ChromeOS Flex on a no-longer-supported old Chromebook like Google’s OG Chromebook Pixel (still a beautiful machine)? Sure there’s that pesky hardware screw, but that’s only moderately difficult to remove.