Intel is well placed to make the leap to mobile

Intel is well placed to make the leap to mobile

Summary: It's far too early to rule Intel out of the mobile game. The company's success with 14-nanometer architecture could give it the power it needs to elbow aside the likes of ARM and Nvidia.

(Source: Intel)

Intel managed to dominated the PC landscape for decades, but when it comes to the mobile market, the Santa Clara-based chip giant is having a harder time elbowing the competition aside. But it's far too early to write off the company just yet.

There's little doubt that Intel has been slow off the mark when it comes to embracing the post-PC world. Whether this is because it was holding out hope for the PC era to continue for a few decades more, or whether it has been biding its time, waiting for the right opportunity, we don't know. Whatever the reason, you're far more likely to find ARM or Nvidia silicon inside mobile devices than you are hardware.

But all that could change.

First, Intel has been making great strides in miniaturizing its architecture, and next year the company plans to push out processors based on a 14-nanometer architecture. This is hugely significant because it will give the company a two to three year lead over its manufacturing rival TSMC. This die shrinking means that Intel can deliver greater performance without sacrificing battery life.

14-nanometer architecture could allow Intel to build quad-core processors with a TDP as low as 1 to 2 watts. In the post-PC world, it is not GHz that matters, but performance-per-watt, and Intel is positioning itself to be able to trump the competition.

Another advantage for Intel is that Google's Android operating system is nowhere near as architecture-dependent as PC operating systems. This means that while the likes of Nvidia and ARM have been able to get into the game early, there's nothing stopping Intel from leveraging its x86 architecture inside mobile devices.

Topics: Intel, Hardware, Processors, Smartphones, Tablets

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • I'm not convinced

    ARM chips are down in the 20ish nm range now, not that far behind. intel is just getting smaller, not changing actual efficiency of the architecture that much.
    • ARM hasn't done much better...

      "Getting smaller" may really be all it takes to jump ahead of ARM in the mobile market though. Intel has had plenty of "more efficient" designs snapping at x86's heels (IBM/Motorola, SUN, DEC), yet none of them have been able to scale as well as x86. ARM is showing signs of the same potential problems in ramping up performance.

      ARM's biggest strength early on in the mobile game has been its low power consumption, which intel is working hard (and seeing success - again) in matching. It's second big strength is that it is a living platform, which allows manufacturers to roll-their own SoC designs. This may quickly become a non-strength as more and more platforms call for standardized designs (closer to what intel has now) and cookie-cutter architecture.
    • Efficiency

      At this point, an atom tablet running Windows 8 gets nearly the same battery life as an iPad.

      They're obviously getting better with efficiency.
      Michael Alan Goff
    • They have both a better technology and significant reduction in size.

      Intel's gates allow them to be shut off (way reduced leakage). The difference between 22nm and 14 nm is huge percentage-wise. It would be the equivalent of going from 65nm to 41nm. This is about a 3 year lead.

      Keep in mind that Haswell FPU can achieve nearly 500 GFlops or performance. And it halved the wattage required.

      Right now Intel has a serious performance advantage and yet it is cutting power.
      • one slight detail

        Those new fabs cost a lot of money. An awful lot of money.

        An fab more or less has limited productivity in terms of number of wafers it can output. So far, Intel was able to command huge prices for it's CPUs, especially the server CPUs. Their math more or less was working because by shrinking the dies, they produce more chips per wafer -- and get more money.

        But the ARM mobile CPUs have another feature: they are damn cheap. Intel has to produce way more smaller (Atom?) chips, in order to compensate for one of the more expensive Core or Xeon chips. Here is how it gets complicated for Intel: those new smaller and smaller process fabs are getting ridiculously expensive and it's simply not worth at one point, if you can't run them at full capacity -- for expensive chips.

        So Intel has to compete with the ARM vendors on all parameters: power consumption, performance and cost. While they are good at power and performance, Intel have trouble with the cost.
        • AMD vs. Intel has shown both can compete...

          Fabs cost money for any type of processor, and any major redesign means an expensive retooling of those plants. I don't see how ARM processors can continue to evolve without similar investments in fabrication technologies.

          Intel will bare the cost of new fabs themselves, but that gives them the ability to innovate faster and control the design process. There aren't many single ARM manufacturers willing to do the same, so any large-scale ARM fabs are likely going to have to be a collaborative effort between multiple ARM licensees - a process that inevitably results in compromise in the technology and a slowing of the innovative process.

          AMD has always been a "price" check on intel anyway, and intel had no problem surviving (and inevitably beating) them, by still offering better technology at a reasonable price.

          The problem with competing only on price is that it is always a temporary advantage - a company with superior products, engineering, innovation, etc. can pretty easily run a promotion to undercut the competition - even if just temporarily. It is not so easy for the price leader to improve on any of the other competitive advantages in their market.
          • Re: Fabs cost money for any type of processor

            The simpler ARM architecture can be implemented efficiently on older, cheaper fabs.
        • Intel can compete with arm on cost no problem.

          Theyve been making giant strides in moving to hugh wafers and getting high yields from them so the cost per chip is significantly declining at a rate arm is nowhere near matching. When the move to 14nm is done they'll have a huge advantage in production cost as well as power and they be increasing the huge perf advantage they already have. While arm keeps throwing more cores at the problem and still lagging horribly on perf intel is quite far along in moving forward to 10nm. 5 years from now only the bottom end emerging market devices will by on arm. All high end phones and tablets will be on intel
          Johnny Vegas
          • Re: so the cost per chip is significantly declining

            Why are Intel chips so expensive, then?
  • Performance, Power Consumption, Cost: Pick Any Two

    ARM is thrashing Intel on performance per watt and unit cost. Intel can match ARM on two of those characteristics at once, but it cannot manage all three. The x86 architecture is just too full of backward-compatibility baggage that is simply irrelevant to most of the computing world*. Sure, Intel has access to more advanced fab technology, but that is also more expensive fab technology, which is going to make your cheap-as-ARM chips even less profitable. Intel has a lot of money to throw at the problem, but how is it going to make any money back?

    *That's right, the PC is now a minority of the computing world.
    • The PC has evolved, and now we're referring to them with the nonsensical

      term of "post PC".

      Fact is that, what the smartphones and tablets do now, is what the PCs of the past were capable of doing, except that, those newer PCs are smaller and with connectivity almost everywhere.

      Also, when it comes to cost, Intel can compete, even if it has to cut the prices of its chips to bring them to competitive levels, and, the bottom line will be improved vial the large-scale sales which those chips will bring. The ones that should be running scared are the ARM chips manufacturers, because they won't be able to compete with the performance of the Intel mobile chips, which will, in fact, offer very good battery life, and will be cost-effective and priced to compete with the ARM processors. Imagine a 64 bit Atom going up against a 64 bit ARM processor. No contest, since the x86 platform is inherently superior in performance, and it might require 2 or 4 ARM chips to match the performance of one single Intel Atom chip.

      But, hey, who knows, and the ARM people might decide to skip the 64 bit architecture and jump to the 128 bit chips in order to stay ahead of Intel's chips. Doubt it, but, the ARM people better have something up their sleeves.
    • Another ignorant "post PC" disciple

      Most people who have a smartphone or a tablet also have a laptop or a desktop.

      It isn't an either/or type of thing.
      Michael Alan Goff