Intel cancels Broxton and SoFIA smartphone SoCs

Intel has had a rough go of the last couple of weeks. Following its April 19 earnings report, Intel announced a plan to lay off 12,000 workers, representing about 11% of its workforce. The company also said it would be re-aligning its operations as it transitions from "a PC company to one that powers the cloud and billions of smart, connected computing devices." We now have word about the product lines under the axe, as well as some clarification of those that aren’t, courtesy of a Forbes.com report.

According to Patrick Moorhead from Forbes, the biggest takeaway from his conversations with Intel is that the company is continuing its focus on the products that TR gerbils will be most concerned with. Moorhead says the company’s bread-and-butter processor lines are not going to feel the cuts very much, as Intel plans to continue focusing on these high-profit products despite slowing growth. This includes desktop and server CPUs, as well as low-power PC parts in the Atom family.

In a statement to Anandtech, however, Intel confirmed that its Broxton SoC is one of the up-and-coming products that's getting the axe. This tablet and smartphone chip would have replaced the Moorefield line of smartphone SoCs released in 2015, most prominently in the Asus Zenfone. ARM processors are playing with home-court advantage in low-power moderate-performance devices, and have put up fierce competition in this space. It is important to note that the cancellation of Broxton doesn’t affect the relatively higher-power Atoms in the Cherry Trail family, Anandtech reports. Those parts power devices like Microsoft's Surface 3 tablet.

Anandtech also got confirmation that Intel’s SoFIA (Smart or Feature Phone with Intel Architecture) SoC is also getting the axe. SoFIA is a heavily integrated x86 SoC that included a 3G radio. It was intended to compete with very low-cost SoCs made for entry-level smartphones and feature phones. Those ecosystems would have been a new market space for x86, and an unusual move for Intel. Even more unusual for Intel, these SoCs would have been fabricated not by Intel in its own fabs, but by TSMC.

Not all of Intel's lowest-power parts are getting cut, though. Forbes reports that the company has reaffirmed its dedication to its Internet of Things efforts. While PC and data center processors are a steady profit center for the company, growth in this space has been slowing for quite some time. Intel told Forbes that it sees the budding IoT market as a potential growth space for a few reasons. The company says the IoT lacks established players, unlike the smartphone and feature phone markets. Intel also figures it has a platform advantage in this space, thanks to the ubiquity of x86 PCs as endpoints and servers.

All told, this news sounds like a silver lining for enthusiasts. While job and product cuts are never good news, we're cautiously optimistic to see Intel's continued commitment to the server and PC CPU space.

Comments closed
    • lilbuddhaman
    • 4 years ago

    But don’t worry, they got their diversity quotas all taken care of!

    • ronch
    • 4 years ago

    When the world’s biggest chipmaker cancels its plans in the mobile arena, it just tells you a lot about the state of mobile, doesn’t it? Back several years ago everyone was scrambling to find the next big battlefield, which was obviously mobile. But just like any market that everyone swarms over to because it’s ‘hot’, the mobile market plateaued, and relatively quickly, much more quickly than the PC industry which saw growth for many years before catching its breath.

    Still, for the few mobile SoC chipmakers that made it, congrats!!! Margins are thinner than a cheap razor and competing in the market is akin to being on a small lifeboat fending off a couple dozen sharks, but you guys made it! Congrats for making a quick buck!!! You da man!!!

    • the
    • 4 years ago

    The interesting thing is that some SoFIA chips were already shipping in Asian markets:
    [url<]http://www.buychinaphone.com/product.asp?/Chuwi-Vi7-3G-Tablet-PC-Android-5.1-Intel-SoFIA-AtomX3-3G-R-Quad-Core-1GB-8GB-7.0-1024*600-2.0MP-Camera-Phone-Call-OTG,1158.html[/url<] Considering that SoFIA was part of a deal with Rockchip and Intel isn't manufacturing them, I really wonder if they can kill off any unreleased models this late in the game. Broxton being cancelled isn't unusual if Intel never got any design wins with it. I guess there last, best hope was a win through MS and they're not aggressively pursuing x86 on the phone side. As for Intel's next move, it wouldn't surprise me if they take a 'can't beat'em so might was join them attitude'. They've flirted with the idea of fabbing chips for Apple before. It wouldn't surprise me to hear those rumors bubble up on news feeds once again.

    • smilingcrow
    • 4 years ago

    I wonder how close Wintel came to agreeing to a Wintel phone?
    W10.1 with a 10nm Core M at a smartphone TDP level that ran desktop x86 applications would have given them a niche that businesses and geeks might have gone for due to having a full fat flavour of Continuum.
    Who here is interested in such a phone/dock combo?

    • blastdoor
    • 4 years ago

    I’m skeptical that they lost in the smartphone market because ARM had an incumbency advantage. I think they lost because they would not accept lower margins on their best CPU designs, and instead tried to compete with a crippled product. If they had competed with their best designs fabbed on their best process it could have been a very different outcome.

    So I don’t follow their logic for why they will have success in the Internet of Things market. Isn’t the “Internet of Things” just a rebranding of the embedded market? And aren’t there incumbents in that market? And aren’t the profit margins razor thin?

    Why should we think this will be different than smartphones?

      • chuckula
      • 4 years ago

      Intel lost because Intel is a chip manufacturer and the only high-margin players that are really left in smartphones are Samsung & Apple, both of whom make exactly zero profits on their chips but make big profits on finished devices.

      Everyone else is a victim (Ti? Nvidia? Broadcomm?), a cut-rate producer (Allwinner) or a company putting out smartphone chips in the hopes of leveraging into non-smartphone markets (Qualcomm, who is perilously close to joining the victim category).

        • blastdoor
        • 4 years ago

        Sure, that’s how the market looks now, years after Intel decided not to compete vigorously in this market. But the market used to look very different.

        Yes, Samsung has a foundry business. But it’s only recently that Samsung has started favoring its own internally designed SOCs for its phones. And sure, Apple has a first-rate design team now, but that wasn’t the case in 2007. It took Apple years to reach the point they’re at now. And according to rumor, Steve Jobs’ first inclination was to go with Intel. It was only because Intel refused to offer a competitive product that Apple ended up going with ARM.

        As for apple’s profit margins… well, TSMC appears to be doing ok fabbing Apple’s SOCs for them. That profit going to TSMC could have gone to Intel.

        Note, though, that I’m not saying Intel made a bad decision here. Maybe they were right to shun lower margin markets. But what I am saying is that I don’t see how the IOT market is any different — isn’t it also a low margin market with many incumbents? Why does Intel think this experience will be different?

        The only way I think it could be different is if Intel decides to accept lower margins in this market. Perhaps they have decided that they are willing to accept lower margins. If so, then the experience could be different. But if Intel thinks they can pull their punches in this market in the same way they did in the smartphone market, then they will have the same experience.

          • SomeOtherGeek
          • 4 years ago

          But that was his point. If Intel made the whole phone, they might have been a different player. It is the players that do the whole thing now that are doing good. Not the ones that do piece meal.

          So, Intel did the right thing to ax it.

            • blastdoor
            • 4 years ago

            If you think I’m repeating his point then I must not have clearly explained my point (or I misunderstood his), because I am definitely not intending to repeat his point.

            Let me try again.

            The market that exists today is not the only market that could have existed. The one that exists today is, at least in part, a result of choices Intel made in 2007 when Apple asked them to make the SOC for the iPhone and Intel refused. If Intel had made a decision to be competitive in 2007, then the market today would look different. Apple most likely wouldn’t be designing its own SOCs. Intel would be making at least as much money as TSMC making SOCs for Apple (probably more, because Intel would be making money off of designs, not just manufacturing). Yet their margins on the SOCs they would be selling apple would likely be smaller than the margins they make on PC and server products. Smaller, but NOT zero.

            Intel chose not to do that. maybe that was the right choice, but it isn’t obvious.

            He also said nothing — so far as I can tell — to explain why the IOT is any different. How does Intel not end up in the same place in this market? The only way I can see is to decide to accept the lower margins that they refused to accept in 2007. Is that their plan?

            • chuckula
            • 4 years ago

            [quote<]Apple asked them to make the SOC for the iPhone and Intel refused. [/quote<] I doubt it was Apple asking and Intel refusing. It was more Intel not having smartphone SoCs ready to go in 2007 especially considering the very first iteration of Atom that wasn't really a smartphone SoC didn't show up until well into 2008.

            • blastdoor
            • 4 years ago

            Intel did have Xscale /StrongARM

            • the
            • 4 years ago

            It actually was exactly that:

            [url<]http://www.cnet.com/news/intel-ex-ceo-looks-back-at-biggest-blown-call-missing-out-on-the-iphone/[/url<] It was Intel's margin play at game. They wanted a high margin per device and ignored the volume which the iPhone would sell. This is indeed one factor why Intel has lost the ultra mobile war. Baytrail which reportedly was sold under cost was too late in the game to win significant market share.

            • blastdoor
            • 4 years ago

            Thanks for the link!

            Perhaps we can now accept that the premise of my question is correct, and return to my question — why does Intel think IOT will be different? Are they now willing to accept lower margins or not?

            I suspect that only the crickets will answer that question.

            • chuckula
            • 4 years ago

            That’s wishful thinking on the part of Intel’s CEO about “passing” on something that Intel never could have won in the first place, like saying that Bernie “passed on” being president.

            It’s more accurate to say that Intel lost any ability to win the iPhone since Intel was not ready in 2007 to be producing smartphone SoCs. It’s not even a question of the CPU core, they theoretically could have churned out an ARM core, it’s more a question of the type of integration that’s needed for an SoC and that Intel of 2016 could do but Intel of 2007 [or more likely early 2006 before the phone was announced] was in no position to do.

            • blastdoor
            • 4 years ago

            Wow, you think Intel is the analog of Bernie Sanders?? When did you become so anti-Intel? I would have thought the tech analog of Bernie Sanders would be at best Nvidia, but more likely AMD. I’d say Intel is the high tech equivalent of Hillary Clinton (if we were to restrict our attention to Democratic presidential candidates). Maybe not very charismatic, but tough and relentless. (I’d say Apple is the high-tech equivalent of Obama, which works for me since I like both, and I’m guessing works for you too but for different reasons)

            Anyway… Apple would have been developing the iPhone for several years before 2007, so the discussions could have started in 2004 or 2005. Intel had Xscale. iPhone SOCs didn’t include an integrated cellular modem (they still don’t).

            If Apple+Samsung were able to do it, then I have to believe that Apple+Intel could have done it. But maybe I overestimate Intel’s capabilities…

            • chuckula
            • 4 years ago

            I’m not anti or pro anything, I’m factual.

            Further, you’re big schtick that the world would somehow be different if Apple had used Intel to fab a chip instead of TSMC doesn’t really hold much water. Apple has shown that they want control much more than they care about some performance metric. If Intel could theoretically have fabbed Apple’s chips it might have been OK for Intel for a while (and very much not nice today given TSMC’s Q1 results with an 18% revenue plunge and big profit drop) but even that isn’t a certainty. If Apple maintained full control, they just could have dumped Intel for a cheaper foundry and the landscape wouldn’t look any different than it does now. Intel has cutting edge fabs, but they ain’t cheap, and Apple doesn’t need cutting edge to sell products. Furthermore, TSMC is willing to make sacrifices to Apple that Intel doesn’t do for any customer, Apple or not. I think the results of keeping customers at bay are showing with Intel managing to keep things steady during tough economic conditions.

            If Intel was chasing its tail trying to produce Apple’s chips, what sacrifices would have been made to other parts of Intel’s product line? Did Intel miss an opportunity with mobile chips? Sure, but Apple is by no means the reason for the “miss”. Also, Intel has a longer term strategy of getting higher margins, and mobile chips have rapidly lost higher margins for all of the players involved. Once again, there is a long list of victims who were all onboard the ARM train and didn’t even “miss” any opportunities but still couldn’t hack it in the market.

            • blastdoor
            • 4 years ago

            [quote<]Further, you're big schtick that the world would somehow be different if Apple had used Intel to fab a chip instead of TSMC doesn't really hold much water.[/quote<] Of *course* it would be different. It would be hugely different. But it's not obvious that it would be better for Intel, or that Intel made the wrong decision in declining to work with Apple on the iPhone. They would have had to accept lower margins, and maybe that wasn't worth it. But again -- how is the IOT any different? If it didn't make sense for Intel to accept lower margins to do smartphone SOCs, why does it make sense to accept lower margins to put chips in IOT devices? Or do they think they offer something that's so much better than what ARM (or MIPS, or whatever) can offer that >50% margins are justified? My guess is that Intel will discover they can't get the margins they want for IOT and after some hand-ringing will decide to back out of that market. I don't think that's necessarily a bad thing for Intel, though. [quote<]I'm not anti or pro anything, I'm factual.[/quote<] Noted.

            • the
            • 4 years ago

            While not suited for smart phone use at the time, Intel did have SoC plans long before the iPhone launch. There [url=http://www.theregister.co.uk/2007/02/06/forgotten_tech_intel_timna/<]Timna[/url<] on their road maps as early as 1999. [url=https://en.wikipedia.org/wiki/Tolapai<]Tolapai[/url<] SoC was announced in 2007 (though shipped in 2008). While not a full SoC, the first Atoms arrived in 2008 as well. Considering there is a 5 to 3 year lead time for SoC development, the iPhone and these Intel chips were underdevelopment simultaneously. The Intel/Apple discussions were rumored to have started in 2005, enough time to have moved the roadmap around to get a mobile SoC in 2008. Granted this would have delayed the launch of the iPhone to get Intel inside but this time table still might have worked for Apple. In the end, the main thing that prevented Intel from winning the Apple deal is that Intel didn't want to sacrifice their margins.

      • Flatland_Spider
      • 4 years ago

      IoT is more like the PC market in that anyone can start a company, and a lot of those companies will be willing to outsource the chip parts to Intel.

      Phones are patent heavy and regulation heavy with only the really big firms able to create a phone.

        • blastdoor
        • 4 years ago

        While it’s true that anybody can start a PC company, it’s also true that they have very little choice as to the OS or the CPU (assuming they want to actually sell to customers). Except for some small niches, the OS has to be Windows and the CPU has to be x86.

        The IoT strikes me as really just a rebranding of the embedded market (but now with cellular and/or wifi), in which customers really don’t care so much about CPU architecture or OS. How is Intel doing in the embedded market?

          • the
          • 4 years ago

          Intel has their niche of the embedded market that is not likely to go away, but the embedded market is heterogeneous soup of niches with an alphabet of established players.

          The reason why Intel isn’t chosen in most cases isn’t performance or even lower power consumption in some cases. Rather Intel prices themselves out of many use-cases due to the razor thing margins other ARM backed companies offer.

          Intel’s Quark efforts, which were to be behind lost cost/low margin IoT devices, haven’t really gone anywhere. That is really telling.

            • blastdoor
            • 4 years ago

            Thanks — that certainly makes sense.

    • Milo Burke
    • 4 years ago

    I wonder if Cyril’s job is affected. Or is he not affiliated with Intel any longer?

    • chuckula
    • 4 years ago

    Looks like Fermi walked in and Atom split.

    • RedBearArmy
    • 4 years ago

    I guess this means an end to a dream of x64 Surface Phone flagship.

      • Neutronbeam
      • 4 years ago

      Was this your sole, personal dream as opposed to, say, ANYONE else in the world? 🙂

        • RedBearArmy
        • 4 years ago

        Pretty much anyone I ask would welcome such a device. Think of it as a phone with integrated WinX on-the-go.

        x64 full Windows phone could end up as THE only PC you ever need for web browsing; watching cat videos, or playing 2d/browser games. Just x64 and not ARM.

        Small screen devices is a market Intel deliberately misses, and Microsoft just doesn’t grasp.
        Ofcourse it may be a plot to put Atom brand in grave and have CoreM power any future designs.

          • cygnus1
          • 4 years ago

          I would strongly consider such a device, however I’m pretty invested in the Apple ecosystem. Not sure if it would be too little, too late for MS for me.

          Now, what would be nice is if Apple made an iPhone that would do that Continuum magic. Assuming they’re switching to USB Type C ports on coming generations of iPhones, I think it would be a lot more than possible. I don’t know how soon/if they could do iOS on x86 though or if the phone would be fast enough to run full MacOS with an iOS layer for apps. x86 may not even be necessary for Apple to do that kind of functionality, it might be better to offer x86 emulation and run an ARM version of MacOS.

        • the
        • 4 years ago

        Actually being able to run Windows as a VM on your phone (with appropriate IO devices like a BT keyboard/mouse and connected to a HDTV) would have its niche uses.

        Or for that matter, a full desktop x86 distro of Linux too for good measure.

      • Generic
      • 4 years ago

      We’ll just have to wait for i3 on a 7nm process. I want Continuum to live dag nab it! 🙂

    • James296
    • 4 years ago

    [url<]https://youtu.be/e_mPrhwpZ-8[/url<] guess their plan failed.

Pin It on Pinterest

Share This