What did Microsoft gain from releasing the Surface Pro X last year?

62

People within Microsoft seem to treat Panos Panay as some kind of genius, but he seems like a one trick pony, who’s basking in the success of the Surface Pro 3. Apart from that, he seems to have released devices before they’re ready. One of the latest examples seems to be the Surface Pro X (another is the Surface Duo, but at least it has apps). It’s a beautiful device, but terribly underpowered for the price. What did they get by releasing it in 2019? Why not either: A. Go the Amazon route and releases plasticky $200 tablet where people expect poor performance or B. Do what Apple did and release it when ready. Surely when they used the device, they would’ve noticed that the performance was crappy. I just don’t get how he gets away with this.

Comments (73)

73 responses to “What did Microsoft gain from releasing the Surface Pro X last year?”

  1. miamimauler

    The funniest part about the Surface Pro X was watching MS's personal cheerleader Daniel Rubino flailing about in an attempt to defend this weak, disappointing and obscenely priced for what it offers device.


    The criticism even among the MS faithful became so extreme Rubino even posted an article telling everyone to stop saying the Pro X is too expensive.


    It was hilarious...ah, good times. ?


    I personally like and use W10 but the Pro X is just another sad example of normal people only wanting traditional desktop Windows and not these app barren RT type OS's...to this point at least.

    • shark47

      In reply to miamimauler:

      True. That's like every Apple reviewer, though.

    • geoff

      In reply to miamimauler:

      I don't get the hate for Surface Pro X. The benchmarks quoted by others in this thread show that the performance is 'good enough'. The form factor is very good, to the point that reviewers are disappointed that the new Surface Pro devices haven't moved over to the Pro X form factor.

      The pen garage is novel and useful. It does touch and pen, which Apple doesn't do.

      It's a good device.


      No, I don't have one. I'm OK with acknowledging that something is OK, even if I personally don't have one.

      • Paul Thurrott

        It's not hate. It's an honest evaluation. Performance that's good enough vs. Performance that is excellent. Compatibility that is not good enough vs. excellent compatibility. Surface Pro X is not good. It's not even OK. Any normal person who purchased one would run into problems, probably on day one.
      • angusmatheson

        In reply to Geoff:

        i cannot find a review comparing the M1 MacBook Air to the Surface Pro X (or any other windows on Arm device. To quote Paul “Benchmarks mean Dick” - what I want to see if how they both handle opening and using the same common apps, word, chrome, photoshop, games. And how they do with battery life. I see a million intel Mac vs M1 Mac videos. Is it because no one has a WOA device? No one is debating should I get an intel Mac vs M1 Mac. Just get the M1 or wait until new M1 form factors. But windows vs Mac? Those are the real comparisons. Surface laptop 3 vs M1 Mac? What does the surface pro x do well? What does the M1 MacBook Air do well?

    • JE

      In reply to miamimauler:

      Yep. I don’t know why I still have WC bookmarked. It’s pathetic.

  2. jimchamplin

    MS needs to expand the suppliers of ARM chips if they intend to go the distance here. Qualcomm has proven themselves completely capable of tanking it when they’re not actually interested. Android smart watches anyone?

  3. shark47

    I'm still holding out hope for a faster Surface Pro X in 2021. I think the M1 makes it more likely that this isn't just going to be a sideshow for Microsoft. In any case, 2021 is going to prove to be a very important year for Panos and Windows and I'm sure he's aware.


    We'll see.



    • bkkcanuck

      In reply to shark47:

      Let us hope that the Mac M1's will light a fire under Microsoft and partners to have the equivalent devices for Windows (and Linux).

    • Paul Thurrott

      Faster isn't enough. For WOA to make any sense at all, it needs to offer a reasonable level of compatibility, not just with apps but with drivers as well. And while I don't think making the PCs faster is particularly difficult, doing so without sacrificing battery life is the other issue. The Pro X already delivers 25 percent less battery life than previous-gen Snapdragon-based PCs. If the battery life falls to 50 percent and we achieve full compatibility somehow, what did we achieve? Intel already beats that.
      • shark47

        In reply to paul-thurrott:

        Agree and they shouldn't bother releasing a new version until they've figured out both, but they also don't have a lot of time. I can't imagine a world where the choice is between a walled garden that is Mac OS and a browser that masquerades as an OS, that is Chrome OS. We need Windows.


        I still find it amazing that while Microsoft was trying to get the 8cx to perform at Core I5 levels, Apple produced a chip that exceeded I7 performance and I5 battery life.

        • ianw789

          In reply to shark47:

          Agreed. We need Windows. But catching up to the M1 will be tough enough, even without the added burden of maintaining compatibility. Let's assume for now that's not feasible, which I think is a defensible position. Then does Windows just live in a dead-end compatibility-focused hardware ecosystem? At this point I need to rethink my agreement with Paul that compatibility is essential.

          • shark47

            In reply to ianw789:

            I don't think compatibility is that important if a. the performance and battery life are there and b. they have an emulation layer like Rosetta 2. I'm in awe of what Apple has done here. I think Microsoft can get close but only if they sacrifice compatibility.

            • Paul Thurrott

              The opposite is true. Compatibility has to be assumed. Then we talk about performance and battery life. If the compat isn't there, the conversation is over.
              • shark47

                In reply to paul-thurrott:

                Right, but for the average consumer, how much does it matter? If you're boring something for school, to bkcanuck's point, most of the software will work well. (Same with the Mac where Microsoft software runs really well.) Microsoft will have to let go of some of the compatibility if they want to move Windows forward.

                • Paul Thurrott

                  I feel like Microsoft tried that, and will try again with Windows 10 X. But it's the user base that's not moving off of classic desktop apps.
                • wright_is

                  In reply to paul-thurrott:

                  I agree. I think it is because people are used to fully functional applications and they use them day in day out. Offer them castrated functionality with a touch interface, when they are still sitting at a desk with multiple monitors and there is just no interest.

                • wright_is

                  In reply to shark47:

                  Except consumers are a small part of the market. Industry is slow to change - and suppliers to industry are even slower and want to be paid thousands of dollars per line of code changed to work on the new system.

                  Internet Explorer 11 is being retired by Microsoft in January, but our PLC control software still uses IE 11 and ActiveX controls and doesn't work in Edge with IE legacy tabs (we've tried). And the supplier doesn't offer an alternative.

            • wright_is

              In reply to shark47:

              Applications are important. The processor and operating system are only needed to run that software. That is something we seem to have forgeotten over the last 30 years, as the world has pretty much gone all-in on Windows and a small minority using Linux or macOS on the desktop.

              A lot of software, and especially drivers, are Intel and Windows only. You can't switch processor types (Intel to AMD excepted) without the applications, and especially the drivers, going with it. And you can't swap operating system if the software and drivers don't work on the the OS. You need full buy-in from all the developers and hardware manufacturers. And you need them to be benevolent and provide drivers for existing equipment for the new platform.

              Given the history of many manufacturers, and I'm talking industrial and laboratory kit here, where you have to buy a new 6 or 7 figure piece of hardware to get drivers that work with Windows 10, I don't see them providing the drivers for free on WoA or macOS.

              • Paul Thurrott

                Exactly. The flowchart for this buying decision is: Does it run all of the apps and work 100 percent with all of the peripherals that I need? If the answer is no, and it is, you've come to the end of the flowchart. It's just a nonstarter.
              • bkkcanuck

                In reply to wright_is:

                I view it as the 80/20 rule or somewhere there abouts. You want to be able to provide a solution out of the box for around 80% of the people currently running Windows - i.e. You want critical mass. So if/when Microsoft has an equivalent solution - they need to be able to run most of the common software that ordinary users use. I would guess that the majority of people use the browser, email, office software, possibly photo editing, video editing, etc -- and maybe the odd additional piece of software as well.


                Laptops make up about 2/3 of the market by themselves.


                Specialized hardware can remain on legacy versions for much longer (with maybe the exception of that used by the likes of audio/video production).


                You also need the excitement of having the new hardware provide a better user experience for those early adopters.

            • bkkcanuck

              In reply to shark47:

              They do have to have a critical mass of equivalent solutions when it comes to software. Microsoft has a lot of that key software themselves (Office 365 [Word, Excel, Visio, Outlook, Teams, etc.], what they need to be able to fill in is the most important missing pieces that would block a wider adoption and use a carrot/stick approach - i.e. maybe subsidizing certain migrations, and/or threatening that either they provide a solution or Microsoft will have to do it themselves (either directly or more likely through acquisitions). Legacy Enterprise software can be accessible through Terminal software. Partnering with AMD could provide the basis for a transitional ARM processor with x86 helper IP for a Rosetta 2 type solution. It is not going to be easy but I think it is possible if Microsoft sees it of core importance. [With the X-Box MS has acquired many game studios to give them a fighting chance]. Affinity could give them some important apps that would be useful (they have their macOS apps already running on M1).

        • bkkcanuck

          In reply to shark47:

          Except that macOS is not a "walled garden" (iOS is, iPadOS is for now). Why I say iPadOS for now, is I believe the iPadOS was separated from iOS for reasons that it will eventually be more iPadOS/macOS than iOS in the future and at that point the walled garden that exists there will have to come down (can't go the other way since macOS was never a console; and moving it to be a "walled garden" would lead to major issues that Apple will not want to have).

      • paradyne

        In reply to paul-thurrott:

        Due to the low-level and performance sensitive nature of drivers, I really don't think translations will work well there, they really need device manufacturers to re-compile for it.


        I remember it taking quite a few years for that to happen when x64 came along, which also needed native compilation for drivers. And so it will be again.


        At the moment, Microsoft has only a carrot, It's Apple who have the big stick as there just won't be any x64 based Macs soon.

        • wright_is

          In reply to paradyne:

          I agree, except that a lot of the code can be written in assembler. If that is the case, it is a complete re-write.

          • paradyne

            In reply to wright_is:

            Fortunately the current Windows is still just the latest version of Windows NT underneath and that was designed to be cross architecture. Originally there was x86, MIPS, DEC Alpha, and Intel Itanic. To make that work the drivers were mostly written in portable C and talked to the hardware via an abstraction layer called the HAL (but there was no pod bay door driver). For most of them it shouldn't be too hard, more a question of willingness to dig out the old code and get on with it.

  4. basic sandbox

    In reply to lvthunder: Qualcomm is the right partner. Microsoft is too slow on the software side.


    • bkkcanuck

      In reply to basic sandbox:

      I don't think Qualcomm is the right partner, I do not think they are hungry enough and have become complacent in areas that they view non-core businesses.


      I have heard recent rumours that AMD has resurrected their shelved ARM processor (don't know if it is aimed at servers or mobile devices). Due to limited resources when Lisa Sue took over she more or less shelved it while she focused on returning their x86 and video graphics business to health (as well as returning their finances to health). AMD is in a much better position now so maybe they decided to reactivate the ARM project... I hope so....


      I had thought that AMD gave up their architectural license to ARM, but as far as I can tell -- it seems they still have it... maybe the project was never completely shelved but in a longer period of R&D (active enough so that development was not lost, but not active enough to be a drain on limited R&D dollars).


      They could be in a unique position for transition silicon in that they have both x86 IP and an ARM license.

    • wright_is

      In reply to basic sandbox:

      No, Qualcomm gave up on optimizing the ARM design they were licensing about 5 years ago. Now they are more-or-less reference designs, the work that Qualcomm had put into optimizing their chips, Kyro, was just thrown on the scrapheap.

  5. peterc

    The surface pro X is a great piece of hardware, its excellent in my opinion. It lacks a suitable OS for it. Square peg in a round hole situation. Stick android and some heavy MS EDGE and 365 app integration into it, or Win10X, or basically something light, touch optimised and an app store and away you go... Full Fat Windows isnt it.

    • shark47

      In reply to peterc:

      If Windows 10X flops, which is indeed possible given their track record (also, maybe they should call it Edge OS), they may release a Chrome OS vor Android version of the device, which would be interesting. Although, Windows 10X is for low cost devices so a Lumia style plasticky device would be perfect to showcase it.

  6. Daishi

    I think the problem Microsoft has is highlighted by the recent reports of developers getting Windows on ARM running on the M1 Macs. With the obvious caveat about how much/little benchmarks alone can tell us, the Geekbench scores they were showing off (1288/5449 vs 749/3014 for the Pro X) make it pretty clear that the problem is the chips they have available for them to use. And no amount of “waiting until it’s ready” would move the needle on that because Qualcomm couldn’t careless about Windows on ARM and without Microsoft pushing it forward there’s no chance they’d bother making anything for it. (It genuinely wouldn’t surprise me to learn that Microsoft had had to cover the entire development cost of the 8CX)


    At this point it feels like the best hope the Windows PC industry has is that, having caught up to and in many ways surpassed Intel, AMD will now focus on catching Apple. Unfortunately that probably means we’re 18 months or more from getting the first products out of that work.

    • Vladimir Carli

      In reply to Daishi:


      so, just to put things in perspective, Microsoft went on the ARM route to put pressure on intel. But, to do so, they picked a partner that is worse than intel. Good move...

      When they realized that the performance of the pro x was so crappy, they could have made it an intel laptop. It would have probably been a huge success. Instead they went this arm route that is turning to be every day more embarrassing

      • wright_is

        In reply to Vladimir:

        If they had made it an Intel laptop, that wouldn't have put any pressure on Intel.

        They wanted to put pressure on Intel, without actually pissing Intel off, too much.

        • Vladimir Carli

          In reply to wright_is:


          Sure, but by releasing a terrible product they actually reinforced Intel's position. They basically told the world that the intel replacement was a joke. Moreover they damaged their brand. The message was that the innovation Microsoft is able to release in the PC world is slow, incompatible, bad battery life and terribly expensive.

          • shark47

            In reply to Vladimir:

            One could argue that Apple has put more pressure on Intel and AMD than anything Microsoft did. Maybe they could've saved millions if they'd waited.


            Also, it might have made sense to stay focused instead of going down rabbit holes like Neo, Duo, etc. The Surface design team is extremely talented and their talent is being wasted on failed products. Just make something that works.

            • james.h.robinson

              In reply to shark47:

              The Mac is a niche player in the PC world, especially in the enterprise. In addition, Apple doesn't allow any Apple Silicon in non-Apple products. The only "pressure" Apple put on Intel/AMD is symbolic.


              Also, the Surface team already makes products that work. That's why they bring in billions of dollars per quarter.


              It's better to experiment and fail than to do nothing.

  7. SWCetacean

    I think it was mostly as an "aspirational" sort of hardware release. That's why they were focusing on LTE and connected standby and such. Release something with the capabilities and form factor, and work on performance and ecosystem later. Pretty much a teaser of what could come.

    As for why they didn't wait to release, I have no special knowledge. Possibly they didn't know what Arm's plans were for higher-performance chips (the A78 and Cortex-X1 weren't announced until earlier this year). Until Apple's M1 released, Qualcomm's 8cx was the only Arm chip with laptop-capable performance.

    • wright_is

      In reply to SWCetacean:

      But LTE is hardly something new. Most of our business laptops have had 3G / 4G modems in them for the last 15+ years.

      The whole project is more a call to get Intel to get their rear back in gear and start actually producing something modern and competitive. The problem is, Qualcomm are the wrong partner for that. They used to take the ARM specs and heavily modify them to make optimized chips, but they canned that several years back and just pump out "reference" designs these days and those are aimed at the smartphone and tablet market, not the mobile PC market; which means they use less power, but are also slower than mobile or desktop chips. Given the "fringe" market that Windows on Snapdragon makes up, it is also not something that either wanted to invest heavily into making - a competitive mobile or desktop PC chip would take 3 - 5 years of lead time and a huge investment by both companies, in return for a couple of million sales, at most, there was just no RoI on such a project.

      Apple, on the other hand, have the capital and the will to see this through. They have set everything on ARM, going forward, so any investment they make is going to see returns in the hundreds of millions of units per year (phone, tablet and PC), as they can use optimizations in one area in the others, therefore spreading the R&D costs.

      Even if Microsoft went all-in on ARM, going forward, there is just too much legacy software that is finicky on even the latest Intel chipsets, let alone being emulated in ARM, that their majority customer (corporate sales) wouldn't touch it with a barge pole for several years, until they can be certain 100% of their software runs natively.

      • shameer_mulji

        In reply to wright_is:

        "The whole project is more a call to get Intel to get their rear back in gear and start actually producing something modern and competitive"


        I honestly think whatever Apple is doing with their custom Mac SoC's and what AMD is doing with their Ryzen series of processors will do a lot more in accomplishing that than anything Qualcomm does. MS is probably better off switching their Surface devices to AMD than they are going ARM.

  8. paradyne

    There's more to a device than just cpu performance, especially if it's one you want to move around easily. I have one on my desk here, connected to a 4k monitor and Edge/Mail/OneNote/Office/Teams etc. (the things I mostly actually use) all run great on it, in complete silence with no heat.


    Also, the only way they could get people to bother compiling their apps for ARM was to have a first party device on sale, people just wouldn't take it seriously otherwise. It's starting to work. Apple's move helps. Now they just need some Qualcomm (or another) to make a decent chip with M1+ performance levels.


    I think it's possible that in a few more years, Intel could just be left as a legacy system if they don't really push ahead with faster chips that could work in a Pro X type device. Everything new and exciting will be ARM (or RISC-V?).

  9. red.radar

    In reply to lvthunder:


    There is a large set of partners that fabricate Arm processors.


    • Samsung
    • MediaTek
    • ST
    • Texas Instruments
    • Microchip / Atmel
    • NXP
    • Nvidia


    Granted some of those suiters have competency in different market segments but I imagine Microsoft didn't try real hard and just did their technical research through Marketwatch. And a few Project managers and Marketing folks went " well geeze my phone seems fast lets try that guy ". And they probably wanted it fast rather than right so .... they went with the biggest gorilla in the room and threw money at them till they said yes. Regardless if they were genuinely interested. And considering how tight R/D resources are these days I bet the B team was put on the project to ensure the bread and butter products made release on time.


    I am surprised they didn't partner with a company who is trying to put Arm in the datacenter that seemed like a better place to start than lets build up a phone processor.




  10. illuminated

    I see a tendency to over--dramatize and panic when the first attempt fails or competition makes something better. So Microsoft released Surface X in 2019, Apple released M1 in 2020. That is a year or more of progress. Let's see what performance the next Surface X chip will have. It would have been absolutely embarrassing if Apple's chip performance would have been worse. They had at least extra year, in-house chip design team, simpler OS, less hardware dependencies and no backward compatibility headaches.


  11. james.h.robinson

    In reply to lvthunder:

    Do you think HP's customers (primarily organizations) care THAT much?

  12. ngc224

    The praise of Panos has always seemed very forced to me too. I’m sure he’s a nice guy and everything, but he’s not getting it done. I think Microsoft tried to create their own Jony Ive, but failed.


Leave a Reply