AMD cutting 7 percent of its workforce and it gave a lower-than-expected forecast

Discussion in 'Graphics and Semiconductor Industry' started by pharma, Oct 16, 2014.

  1. pharma

    Veteran Regular

    Mar 29, 2004
    Likes Received:
  2. Grall

    Grall Invisible Member

    Apr 14, 2002
    Likes Received:
    La-la land
    What kind of people can AMD really afford to fire these days - not engineers, one would think, and surely their administration organization has to be pretty damn lean these days considering they've been running at a loss for years now?
  3. Alexko

    Veteran Subscriber

    Aug 31, 2009
    Likes Received:
    They could probably survive without the people in charge of tying CPUs to weather balloons.
  4. 3dilettante

    Legend Alpha

    Sep 15, 2003
    Likes Received:
    Well within 3d
    This news does seem to fit in the doom and gloom thread.

    As far as who AMD can afford to let go, I'm not sure if that set is large enough to encompass the number it is likely to let go in the upcoming contraction, and I hear a ring of truth to predictions that AMD is likely to have another round after that.
    There are reports of AMD outsourcing its chipset design work to ASMedia, and a transfer of IP and 150 or so engineers to Synopsis in exchange for access to more tools and IP.

    It does look like AMD has already been trying to extract trade value from its players before it cuts them outright.
    The payoff of the transfer as given in the press release is apparently up to $450 million in cost efficiencies.
    The cost is that some chunk of what underpins AMD's R&D and interface IP will in a little while cease to be a differentiator for AMD and might be used to deny it future customers. If the engineers involved were primary contributors to this, it sounds like a big chunk of things involved in high-performance off-chip IO is no longer going to be something AMD can uniquely offer.

    Perhaps the argument could be made that the expanded access to multiple advance node IP and toolsets was a bigger win, since AMD needs to hedge its bets more on being able to field customer requests across a more diverse terrain that it could ever afford to cross on its own. At the same time, the need to cater to a generic client set means bespoke/superior solutions would be counterproductive.

    If AMD is fully drinking the 2.5D integration colored-sugar-water-drink, it might be counting on a more limited subset of interface IP for on-package or stacked communication and less demanding tech for the remainder. With the higher levels of integration, much of it would based on what the foundry is offering anyway.
    Intel and Micron could come into the picture if AMD gives up on controlling its own memory destiny and goes HMC, and that doesn't seem like an impossible next step given everything else it's trying to leverage... customizable... synergistic... efficiencies... execution... uh... buzzword from.

    The negative interpretation is that AMD could trade some of the crown jewels and a portion of its brain trust because the competitive quality of the former and the quantity of the latter are only going down anyway.
    I'd like to think AMD will still have something compelling to integrate once it reaches its ideal level of customization ability, as well as something well-improved as it is well-leveraged.

Share This Page

  • About Us

    Beyond3D has been around for over a decade and prides itself on being the best place on the web for in-depth, technically-driven discussion and analysis of 3D graphics hardware. If you love pixels and transistors, you've come to the right place!

    Beyond3D is proudly published by GPU Tools Ltd.