Surface 'Centaurus' is another dual-screen Microsoft PC with Windows Core OS

Photo credit: The Verge

Microsoft is working closely with Intel on a new dual-screen Surface device powered by Windows Core OS that's similar to Intel's Copper Harbor prototype that was revealed earlier this year. Codenamed "Centaurus," this device is akin to Microsoft's canceled Courier project, which saw the company conceptualize the idea of a digital journal in 2010. Centaurus marks the second dual-screen device we believe Microsoft is currently working on internally, the first of which is codenamed "Andromeda."

What we know about Surface 'Centaurus'

According to my sources, Microsoft has been working on Centaurus for about a year now and is hoping to announce the device in the fall of 2019. It is supposed to be the first of what Microsoft calls a dual-screen 2-in-1, a new device form factor for Windows that utilizes two screens for productivity, inking, and more.

Like it did with the Surface Pro, Microsoft is building Centaurus to show hardware makers what can be done with this new form factor. These devices are adaptable and can transform into a tablet, laptop, digital book, a bigger tablet, or somewhere in between. Since Microsoft is working closely with Intel, it should come as no surprise to hear that an Intel processor powers Centaurus. This means Centaurus will be able to run x86 apps.

I'm told that Centaurus runs a flavor of Windows Core OS, Microsoft's new modern version of Windows 10. It runs an experience that has been designed from the ground up for dual-screen devices, and as such, provides a tailored experience that takes advantage of the dual-screen setup, that can also adapt the user experience to fit other orientations the device may be usable in.

Photo credit: PC World

While Windows Core OS can run legacy Win32 programs, it can only do so if they are containerized or are from the Store. As such, Microsoft has been focused on getting as many Win32 programs into the Microsoft Store as possible during the last couple of years. This means there will be a library of Win32 programs for users to take advantage of on Windows Core OS devices like Centaurus.

Since Windows Core OS is adaptable, the user experience can change on the fly. For example, if Centaurus is being used in a tablet orientation, you can fold it into a laptop position, and the OS will adjust to provide an experience akin to a laptop. This would make one screen a keyboard and trackpad and the other screen a familiar desktop with a taskbar along the bottom and windowed apps.

What does Surface 'Centaurus' mean for Andromeda?

Surface Andromeda is believed to be a pocketable dual-screen phone, and reports from earlier this year had suggested that Microsoft was scrapping the pocketable phone idea in favor of a larger tablet form factor. This larger version is Centaurus, but we hear that Andromeda hasn't been scrapped.

I'm told that Microsoft has put Andromeda on hold and is prioritizing the release of Centaurus instead. That's all in the name of ensuring that Andromeda has a successful launch when Microsoft finally decides to re-enter the mobile market. The problem with Andromeda is that when it's being used as a phone, it's too small to take any real advantage of Win32 programs, and there aren't enough Progressive Web Apps (PWAs) or Universal Windows Platform (UWP) apps in the store for people to warrant buying one as a phone.

If Andromeda were to ship in the next year, the device would launch with no apps, just like Windows Phone did, and it would more or less be dead on arrival. Microsoft is pushing hard on PWAs to help solve this, but it's going to be a while before any of that takes hold. In the meantime, Microsoft will try to kickstart this new form factor by releasing a bigger version, one that isn't a phone and doesn't depend on UWP apps to be useful. And it will try to convince OEMs to do the same.

If this form factor proves to be popular, the idea of a pocketable version becomes much more appetizing. Microsoft's goal is to give the larger dual-screen 2-in-1 devices a little time to breathe, push developers to build UWP apps that take advantage of the dual-screen setup, and give PWAs more time to mature. Then perhaps in a couple of years, a device like Andromeda will make much more sense.

As far as I know, Microsoft doesn't have a release window for Andromeda right now. It's still being worked on internally, and Microsoft really does want to ship it eventually, but it's waiting to see how Centaurus is received first. I'm told Microsoft is also considering (but hasn't committed to) the use of Android apps on Andromeda as another way to fill the void of apps when being used as a phone, but that's something I'm still digging into at this time.

For now, Centaurus looks like it's going to be Microsoft's first foldable PC to launch during the next year. It will be the first new Surface form-factor since the Surface Laptop of 2017 and the first new form factor device for Windows since the introduction of the 2-in-1 with the Surface Pro. It will also be one of Microsoft's first Windows Core OS-powered devices to ship publicly, as Surface Hub 2X isn't expected until the 2020 timeframe. Of course, all of these dates can and will change, as nothing is set-in-stone when things are still being developed.

Zac Bowden
Senior Editor

Zac Bowden is a Senior Editor at Windows Central. Bringing you exclusive coverage into the world of Windows 10 on PCs, tablets, phones, and more. Also an avid collector of rare Microsoft prototype devices! Keep in touch on Twitter: @zacbowden.

164 Comments
  • "I'm told Microsoft is also considering (but hasn't committed to) the use of Android apps on Andromeda as another way to fill the void of apps when being used as a phone, but that's something I'm still digging into at this time."
    Please, no. I don't want crappy hangdroid apps on Andromeda.
    I'd rather rely on 3rd party UWP apps.
  • I think you might be in the minority on this one. I too would prefer UWP only, but if Microsoft can nail a smooth emulated Android app experience, I don't see why we can't have that too. UWP on its own means Andromeda can't exist, unfortunately.
  • I agree with Zac here. UWP would be preferred. I prefer UWP over PWA. But when UWP isn't an option, it's nice to have PWA. As long as Microsoft can commit to UWP for their own apps, I can accept the option of developers simply loading their Android apps into the app store. I would also appreciate it if store listings show if it's an Android port, UWP, or PWA. That way if someone would "rather rely on 3rd party UWP apps" they can see it in the store and make their own decision.
  • For me, it depends on the App. I don't see a reason to have Whattsapp as a PWA. All Apps that rely mostly on a server backend, make sense as PWA. Everything that needs a more powerful client or works "independent", should be an UWP.
  • Agreed. If we had Android apps on Windows "phone", I would probably not be using an Android phone right now.
  • Exactly right.. the market and developers have spoken. You can run Android apps on Windows, or run Android apps on Android, or buy an iPhone. There is no other option for a small/mobile device.
  • I wouldn't mind Android apps as long as they are curated through the MS store. Say what you will about Apple, but the apps available through the AppStore are generally good quality and malicious apps are usually weeded out. A decent selection of popular Android apps would be welcome in an attempt to avoid the "app gap" that Windows Mobile suffered from.
  • I thought they might try and partner with Amazon getting either the Amazon app store on its own or better for Microsft integrate it invisibly into their own store. It will be difficult for Microsoft to get developers to submit their apps if they aren't interested in Windows so partnering with large app stores outside of google could work for them. Obviosuly Amazon need some way to make a profit from this though.
  • Before suggesting the Amazon app store, buy yourself a Kindle Fire and try using the app store from that. Stuff is either missing or out of date. Very disappointing.
  • The problem is. If Android apps run on the device, why develop native apps from a developer's point of view? Also if Android apps run emulated, it only means that they will do the same thing, (or less) but slower nd consuming more battery due to emulation. It will be just like Facebook and Messenger apps for WM10, where they are super-slow and super-huge due to having to run an iOS subsystem as an additional layer. The outcome is that they are eventually the same app as on iOS, only slower and eating up more RAM for the same stuff. Running Android apps on the device would mean nothing more to Microsoft IMHO than having an Android phone with Microsoft products on it. It can only mean less. Long story short, I don't see how a Surface device having to run Android apps is worse than a Surface device having to run Android. I could be wrong. But I don't yet see a point in this move.
  • Android apps don't necessarily run emulated, and might not be slower than on Android smartphones.
    Windows 10 has the Linux subsystem, which isn't emulation or virtualization, but a compatible subsystem side-by-side with Win32, on top of the NT kernel.
    It actually originated in the efforts to run Android apps on Windows Phone, as Android is running on a Linux foundation.
    Seeing Linux binaries can now run on LXSS (the Linux-compatible NT subsystem) as fast and even sometimes faster than on real Linux (technically, CPU-intensive are often faster, while filesystem access requires some extra steps to run over NTFS and gets slower), this means we could get an ARM version of LXSS that can run the Android system as fast as a native Android smartphone. The only real problem is we would probably get many of the bad habits of Android as well, such as a bunch of processes running in the background all the time. But a curated selection of apps through the Microsoft Store, as well as some by-design limitations, such as allowing background processes on an app-by-app basis, could make it perfectly fine. If done right, Windows 10 could end up being better at running Android apps than Android itself, with the added benefit of running UWP and Win32.
    As to getting developers to jump to UWP, Windows Phone showed a new pocketable platform that forces developers to provide UWP apps doesn't work. But they have two other solutions they can take advantage of to still make it happen.
    One is their development platform. Xamarin, Unity and other runtime layers can move developers to a more abstract platform than gets compiled to both Andoid and UWP, developers can then jump to the UWP platform with a simple new compilation target.
    Second is their focus on mixed reality. Since that is a new platform and new user experience paradigm, if Microsoft succeeds with their Windows Mixed Reality platform, which only supports UWP, Win32 and WebVR, they could make developers build apps that scale back to 2D UWP when running on other Windows devices.
    Basically apps built from the ground up to support both mixed reality and 2D could be UWP from the start.
  • The only problem is that Linux apps actually exist, while Android apps don't exist. What you have in reality are Google Play apps, so Android emulation brings almost 0 apps to start with. Of course it may make for developers easier to submit the app but the problem is that by now developers know that Microsoft rarely paid them enough to justify even the trouble of uploading the screenshots.
  • There is the Whole country of China that would disagree with you. There is no play store in China and they have plenty of apps. Europe is heading this way too. We are the only ones stuck with the play store.
  • Actually European ruling is entrenching Google Play - previously you had to bundle all Google services if you wanted Google Play, now you don't have to. So previously there was a remote chance that someone would launch a pure Android if he wanted to ship the phone with non-Google apps. Now only a mad man would launch a phone without Google Play.
  • Europe? Not that i would know of (I'm German). But personally i use a Android phone without google services, getting all my Apps via FDroid and APKPure.
  • @Philippe Majerus
    If Microsoft really embraced this multi-subsystem thing you're talking about and pushed that as a defining feature of Windows, that would probably make me more excited for the platform than I've been in many years. I could see it maybe working in concert with Hyper-V so every subsystem is virtualized separately.
  • I agree. Multi-subsystems is a great plus, not a negative at all. We should just be able to have the apps we need regardless of the tech beneath.
  • Rumors where that APKs ran better on Windows mobile than on Android. They also said that APKs ran better than UWP (Which is supposedly why they killed it). Either way, I don't see a negative here. Apps are what we need. Once you get enough users, native apps will come, assuming UWP gets better performance and integration.
  • You could test Android apps on some Insider builds and it was very bad. Microsoft demonstrated it live and it was buggy then too, so why rumors when there are facts?
  • 1. Who cares what the technology is behind an app, the only important parts is it available and does it work well. No one cares how the code is written. 2. It wasn't "Emulation" they used API redirects, similar to how x86 code works on x64 Windows. you don't really need Emulation unless you changing processors, Java is processor agnostic, like C#. Supposedly Android APKs worked better on Windows Mobile than they did on Android and supposedly better than UWPs too (which is supposedly why they killed it). 3. Facebook IS NOT written for iOS and then emulated elsewhere with an iOS subsystem, it is using ReactNative which is a JavaScript framework, then it hosted on node which is the Chrome Browser engine. Blame node for the slowness.
  • react native is not hosted on node. It uses the javascript engine that is on the OS.
  • It would be much better if Microsoft made it easy to port Android apps to UWP. They've worked hard to create XAML Islands for WPF and WinForms, I'd like to see a little bit of the old "embrace and extend" in regards to Android apps.
  • I submitted that Idea a long time ago. All they would need to do is create a Java interpreter for their Roslyn compiler then use API shims for native APIs. Shouldn't be to difficult for a good team of coders. EDIT. Looks like there is an open source library to compile Java to .Net. http://www.ikvm.net/ While this isn't APK to UWP, it definitely is a proof of concept.
  • While I don't like the idea of Android apps, I don't think there is an alternative. The reason I am using an iPhone and iPad is apps. I haven't bought a Surface because there just aren't any apps available. I don't there is any potential future where Microsoft Store is a viable 3rd alternative to Apple and Google.
  • There are a TON of apps for the Surface... they’re just legacy and not UWP apps. Having said that, few Windows apps are optimized for touch - most of them are simply desktop apps that can be used on a tablet. Also, tablet mode on the Surface is a joke - it’s really just ‘alternate launcher mode’. Curious to see how (if) CoreOS improves the situation.
  • I thought those bridges was the tools for developers to convert their projects to Windows...
  • That depends on which bridge, some are a compatibility layers at the tools level and API shims, while some are runtime environments. Basically Centennial (Win32) is about allowing full access to the Win32 API from apps distributed through the Store, hence the "access all your computer resources" (aka Full trust) autorisation they require to run, as there are no runtime access rights enforced on these APIs. The compiled files are exactly the same as for desktop apps, only their packaging and some runtime redirection (to isolate their filesystem and registry) differs. Islandwood (iOS) is about supporting the Objective-C language and iOS API compatibility shim so iOS code compiled against the bridge libraries end up calling into the Windows API, once compiled with the bridge toolset, they are effectively UWP apps, just using unusual libraries. Astoria (Android) is dead, but lives in the Windows Subsystem for Linux (WSL) and could be brought back to life if needed, benefiting from the years of work on WSL.
    For both Astoria and WSL, the idea is to run binaries unmodified. To achieve this, a complete subsystem that supports the Linux kernel ABI (i.e. system calls) is added on top of the NT kernel, running next to the Win32 subsystem. This has already been done in the past to support UNIX and OS/2 subsystems on Windows NT.
    In this bridge, the complete Linux user-mode is running on top of the LXSS kernel, which is not a Linux emulation or virtualisation, but NT talking the same language as Linux, so user-mode code believes it is running on top of the Linux kernel.
    From there, Windows can run the user-mode environment from Ubuntu, Debian, Suse, Kali, CentOS,... unmodified and even simultaneously side-by-side.
    Running Android apps is basically just a question of running the Android user-mode environment on top of LXSS instead of Linux.
    The only real problems with running Android this way are the way Android lets app run code in the background (bad apps would still behave badly on Windows), and legal rights to run Google binaries on top of LXSS.
  • IMHO, MS would be wise to bring APKs to the Desktop Windows Store, not just for mobile devices. This would start seeding the mobile environment and work out the kinks before a proper mobile platform launch. This would also play into their effort for app mirroring on Android phones on the desktop.
  • I really don't see the point, then. Why not just go all-in for Android at this point? I'm not seeing a benefit of a Frankenstein's monster device. I was never fan of the Andromeda device since it will use untried tech (that will likely NOT survive real use for very long) and be too large to conveniently fit in pockets REAL people have and just be too niche. Depending on the size of the Centaurus device, and the apparent design of being two separate screens instead of a folding one, this would be more convenient for the Dayrunner crowd. I'd still prefer a Surface Pro or Go, because for real work, the size and form factor are a more appropriate platform. I could see, though, the Dayrunner contingent would find a Centaurus form factor more advantageous.
  • Yea, I hear you, I just want a 6" slate device with pen support, a great camera, and Windows with a Mobile Optimized UI/UX. I don't need the folding gimmick.
  • They basically just reinvented the Courier that they stupidly scrapped 10 FREAKING YEARS AGO and they're now thinking that this is innovative in 2019? They would have had something if they released the Courier right before or right after the iPad, but now??? If they're worried about an App gap with Andromeda then why aren't they worried about the same thing with Centaurus? It makes ZERO sense. Who exactly is this for when we already have the Surface line of computers? You can already consume content, ink, and type on any surface device. You can even use two apps side by side and configure how much screen real-estate each app takes up without being constrained to a fixed size... Let's face facts.... The only reason why Microsoft is releasing Centaurus before Andromeda is because they are being LAZY and SAFE. It's easy to use plain windows 10 with a fixed side by side view to work with a Centaurus Intel device. However such a device is entirely useless. Andromeda on the other hand would be useful, people would want a phone with the ability to increase screen real estate or to multi-task, but they wouldn't want to carry around a handicapped laptop or a tablet with a big crease down the middle.... A phone that could be useful for gaming, video editing, photography, etc. would fit a real niche. If Microsoft thinks they can just stumble into a new device category by slapping windows on a tablet with hinges, they have some serious idiots working there these days. Surface was a well thought out device that fit a real need in the market. Windows Phone was a well thought out device that tried too hard to chase Apple when they should have focused mostly on business while that market was being under served. Then they just decided to give up after finally perfecting the mobile experience and hardware. Centaurus looks to be a poorly thought out device with absolutely no focus.... They are going to release Centaurus late (after Samsung releases their fold-able phones) and then after they realize that most people don't give a crap about it, they are going to scrap Andromeda.
  • Yeah, although I'm usually an early adopter, I just can't for the life of me see super value in any thick as a brick device in your pocket. Too heavy to really be considered "pocketable" and that weird crease in the middle is kinda well... ugly. Would much rather see Microsoft do everything in its power to keep our phones in our pockets as much as possible (at least when we're in reach of a laptop, 2 in 1, 10 inch tablet, Desktop, or Surface Studio) by Improving Your Phone to the point where it's a truly seamless, no thought required, pleasurable experience. Screw Apple on Your Phone... I'd drop those closed-system elitists altogether... and go all in on Android.
  • This is spot-on correct, Zac! I said back when too, that having Android apps would not destroy UWP development as was commonly assumed. In reality, it would have allowed the public to buy devices that they knew would work with their banking, airline and other mission-critical apps while showing developers that there were millions of people willing to invest in Microsoft mobile devices. Native apps would naturally follow, especially as tools like Visual Studio caught up and made universal development painless. This is still the only path forward for Microsoft and they should jump on it now!
  • Yea, they should have had a three pronged approach. 1 Allow apps from anywhere run on Windows Mobile (UWP, Android, iOS, Win32, Win64, Linux, etc...), this would have given Windows Mobile the largest library of apps. 2. Make UWP into UAP (Universal App Platform) truly universal so that UWP would run on all major platforms (iOS, Mac, Android, Windows, Windows Mobile, Web) with as little change as possible. This would have made UWP the go to platform to develop for. Platform UNO has made some great strides in this department but don't seem to be getting any support from MS. 3. Fleshed out th