ARM 20 nm Processors Expected to Arrive Next Year

ARM's Simon Segars, general manager of processor and physical IP divisions, said that ARM-based 20 nm SoCs could be appearing in smartphones as early as late 2013. Given the fact that 28 nm chips are still in short supply, the move to 20 nm within 12 - 18 months is rather quick.

Intel is placing big bets on its manufacturing technology to compete with ARM chip makers. The current Medfield 32 nm will see a new 22 nm chip next year, which will be replaced again with a 14 nm SoC in 2014, according to Intel's roadmap.

"The whole industry is focused on moving to the next generation as soon as it's economically viable and technologically achievable," Segars told press at Computex in Taipei. There is no information to back up Segars' prediction, but it is common sense to expect such notes given ARM's huge exposure at this year's Computex as Windows RT arrives and a number of ARM-based Windows notebooks are shown. As ARM's profile grows, however, greater investments in manufacturing are likely and a 20 nm forecast for 2013 may not be so unrealistic.

ARM's revenue has been just $294.9 million for the first two quarters of its current fiscal year, but the IP company said that more than 1.9 billion processors based on its blueprints were shipped globally during the time. 1.1 billion chips went into mobile phones and tablets.

Create a new thread in the US News comments forum about this subject
This thread is closed for comments
27 comments
    Your comment
    Top Comments
  • blazorthon
    process_orIf Atom's gains going to 22nm are as trivial as the gains from Sandy to Ivy, then 22nm Atoms still won't be competitive with Tegra 3, let alone Tegra 6 or whatever will be out then.Process technology is a crutch, if 22nm is this pointless, 14nm is going to be bad times for Intel.


    Intel's 22nm process worked excellently for Ivy Bridge. It lowered power consumption substantially and that was the point of it. Intel skimped on overclocking by replacing the fluxless solder that is usually between the IHS and the CPU die with poor quality paste, but that has nothing to do with the 22nm process itself. Ivy also isn't much faster than Sandy, but Ivy was just a die shrink. Die shrinks don't increase performance much at all. They never do. Shrinking the die gives room for a better architecture. The 22nm tock for Intel will be Haswell. Haswell is supposedly going to have substantial performance improvements like the previous tocks have had and it will do it all on the 22nm node.

    22nm isn't pointless. It did exactly what it was supposed to do (lowered power consumption) and it did it well. 14nm will probably do it again too.
    10
  • Other Comments
  • psikaikai
    Yay competition with the Intel monopoly
    8
  • TheBigTroll
    hopefully, arm will become the new amd
    -5
  • Anonymous
    If Atom's gains going to 22nm are as trivial as the gains from Sandy to Ivy, then 22nm Atoms still won't be competitive with Tegra 3, let alone Tegra 6 or whatever will be out then.

    Process technology is a crutch, if 22nm is this pointless, 14nm is going to be bad times for Intel.
    -2