Microsoft has announced, rather early if you ask me, that the Surface 3 is going to be discontinued and no longer manufactured as of December of this year. It seems like an extreme move, given that the Surface 3 is their low end product and is under no lifecycle commitments, which are probably the reason that the Surface Pro 3 is still being built.
Thurrott's own thoughts on the matter are interesting. Over the past year or so that I've been reading his site, I've seen a lot of big talk about predictions and ideas. In this case, he predicts that Microsoft could kill the non-pro Surface off entirely and replace it with a Surface Phone, which would probably retail for a similar price, and then add a tablet display or lapdock.
With Thurrott's Surface Phone suggestion you end up with something a lot closer to what the original Surface RT was supposed to be, way back in the day, except that it's a tablet or laptop connected to a phone via cables or something like WiGig. It's probably going to cost more, it's probably going to be slower, and this time, instead of compiling ARM verisons of desktop Office, Microsoft is probably going to saddle it with UWP software, which means that when the UWP platform inevitably flops, there will be no more new software forever. It would end up exactly like a version of the original Surface RT except one that cost $900 or so, which is what a Lumia 950 or 950XL and either a lapdock or some basic hardware for continuum use costs.
I like the idea of this product, and HP does too, but I don't think it stands in for the Surface 3, simply due to the cost and performance implications of Windows 10 Mobile on ARM silicon with (probably) less RAM and similar storage, plus with how much more it will cost. Thurrott himself acknowledges in a comment to the article that it's essentially Windows RT again, and that we should all be wary of it. Of course, I've spilled many words on why I love the Surface RT (and by extension to a certain point, the Surface 2) but I can't imagine myself in love with a Surface phone that can't do any of the more esoteric things the Surface RT did (map a drive, windowed remote desktop and/or RemoteApp access, text editor, file system access, SD card, and so on) costs more and is limited to the availability of the app store. If Microsoft kills off the UWP store, desktop Windows 10 computers can still run Win32 programs. The Surface RT can still run the programs that came with it, a Surface Phone would be dead in the water.
I also don't agree with the idea Thurrott seems to have that the Surface 3 is either a bad device or bad at being a Surface. It's nowhere near as high end as the Surface Pro devices, but it's also not outrightly bad (except for the eMMC storage, which is definitely not ideal, and an upgrade to SATA solid state storage is probably now doable in Surface price ranges.) If the idea is that the Surface 3 doesn't fit within Microsoft's Surface product line, my question is firstly what I'm supposed to buy instead (don't answer that) and secondly, why, in a world where Apple was still selling The 101 in its retail stores until last week.
The Surface 3 fills a niche that I think is important. It's got a lot of capability in that it's a tablet that competes with the iPad Air 2 and 9.7-inch iPad Pro. It and its predecessors have had keyboard and multitasking functionality for a bit over three years before the 12-inch iPad Pro came out. It also works well as an entertainment and gaming tablet. As with Windows 8, a lot of core tablet apps are on it. Kindle, Hulu, Netflix, things like that. The kickstand makes it good for setting on a table or near a bed or chair, and the WWAN version would be good for travel.
I've probably mentioned this before, but I have a few different ideas for a successor to the Surface 3. It's theorized that the Atom X7-Z8700 is probably not actually running t "only" 2W, so a Pentium N3700 would provide similar performance (albeit with a slightly less cool name) and the ability to use SATA storage, which could be a big benefit, in terms of performance.
The big changes, physically, I would like to see are the new keyboard mechanism from the newest Type Cover Pro. A version of the keyboard cover with the fingerprint sensor for Windows Hello would be great, but I don't know if it could sell in anywhere enough volume. I'd personally also like to see a new power cover with a big battery in it. The other things I'd like to see are a migration to USB Type C connectors to replace the power, USB, and DisplayPort connectors, or at least the power and DisplayPort connectors, leaving one USB 3.0 Type A port on the system.
I'd carefully save my pennies to get a newer system, because although the performance of the Surface 3 is fine with me, buying new MicroUSB cables all the time is costly and I'm sure the connector in the system will eventually finish dying.
The other potential advantage of a Surface 4 based on the Pentium N3700, or its successor the N4200, is that these chips can also use 8 gigabytes of RAM. Between that and SATA storage, it may be possible to offer further upscale configurations. I would personally be very interested in something, with, say a 256 gigabyte SSD or 8 gigabytes of RAM, to get that much more life out of it as a primary portable computer.
Although, by that point, the inexpensive models of the Surface Pro might start to make more sense. Microsoft, as far as I know, isn't a very high volume PC maker, and I don't know how well the product line could survive crossover like that.
Another suggestion has been to put some kind of "Y" series CPU in the smaller Surface product. Whether it's the "9.7-inch Surface Pro 4/5" or a Surface 4, there are Pentium variants of the Core M CPUs, which would both be a significant performance boost over the Atoms currently in use, and allow for most of the changes I'd like to see.
The Pentium 4405Y should be able to run an Alpine Ridge controller, but I suspect that will be an option (if Microsoft ever bothers with Thunderbolt, after four generations of computers that have failed to have it, even though they had the connector) will be restricted to the Surface Pro family.
The other pie in the sky wish that I have had for the Surface family for a while is Intel wireless. The Realtek solutions that they currently use are okay, but I've seen situations where it slows down after a relatively short distance that even old Intel wireless chips like the Pro/Wireless 5300 covered easily.
With the transition to USB Type C, ideally on both the Surface and Surface Pro/Book families, the opportunity for branded cables and various docking or power adapter solutions is there. In terms of accessories, I'd be entertained by a Microsoft-branded power bank, but I think another power cover might see some success. A power adapter like the one included with the Core M based Intel Compute Sticks that has SUB ports on it would be great, and something like that that further includes, say, an Ethernet connector and one or more display output would be great.
Although, again, there's the question of at what point you should just look at the Surface Pro and accept that within the Surface family, the functionality you want costs more and is bigger. Even with a more powerful Surface, there's the question of how necessary it is to use it docked at home. Desktops that are much more powerful than Surface hardware are inexpensive and easy to get, and file servers, NAS systems, and online storage services mean that you can easily have the same data on multiple computers.
To get back to Thurrott's point, I simply can't see Microsoft trying yet another flagship phone, especially given that the already small Windows Phone user base shrank in the face of the Lumias 950/XL, 650, and 550. This happened for a confluence of reasons, most of which add up to an environment where I just can't see a Surface Phone succeeding, unless Microsoft builds really compelling tablet, laptop, and desktop docking stations. At that point, though, why not tack a battery onto something vaguely shaped like a NUC or the InFocus Kangaroo? Almost nobody will want to buy a Surface Phone to use as an actual phone. As Thurrott proposes it, the Surface phone also loses the biggest advantage of the Surface 3: The fact that you can run Visio, Photoshop, and Visual Studio alongside Netflix, Twitter, Facebook, and Kindle.
I find Thurrott's idea intriguing, but at the end of the day, I think he's misguided in thinking that his particular version of the Surface Phone will save the platform, or somehow be a better product than the Lumias 950 and 950XL were, especially when, at this point, there's not enough advancement in technology for a Surface Phone to be anything else than a dressed up and maybe slightly faster version of the aforementioned Lumia.