Blog Post

AMD Plays Catch Up on 45nm

Today AMD said it has 45 nanometer chips for desktops and servers running in development systems. That’s fantastic, but the chips won’t be in actual devices until the second half of the year, putting AMD’s most advanced chips at least six months behind Intels’s 45 nanometer chips.

Also, one of the touted benefits of 45nm is increased power efficiency, which makes them a good fit for mobile computing. Intel launched its 45nm chips first in a new laptop, but AMD won’t have 45 nm chips in laptops until 2009. By then Intel will be working on 32 nanometer chips, which leads me to wonder how long AMD can play this game.

5 Responses to “AMD Plays Catch Up on 45nm”

  1. Stacey Higginbotham

    @Jesse, the platform strategy is very sexy. I would argue tho, that AMD has made similar pushes beyond the PC market with its acquisition of Alchemy Semiconductor in 2002 and attempts to offer low-power multimedia chips for mobile devices. That didn’t fare well. I do believe AMD has some fabulous designers and innovative thinking, so I’m both hopeful and skeptical when it comes to their Puma and Griffin efforts.

  2. Jesse Kopelman

    One thing AMD is doing right is that their acquisition of ATI has allowed them to introduce platforms where using all AMD components works better than mixing and matching. For example, thanks to Hybrid Crossfire, that low end AMD video card will actually work better on an AMD chipset motherboard than it will on an Intel or Nvidia chipset motherboard. Now, AMD needs to make sure consumers understand this. Also, the push beyond the traditional PC market is something that Intel has struggled with in the past. Their new efforts in this regard have the potential to stretch even their titanic resources. If AMD can keep its head in the game, it should remain a viable competitor to Intel in the traditional PC market. The positions are somewhat like BMW compared to GM, say 20 years ago. Look how well that has played out for the respective parties. Of course, I don’t believe BMW was ever saddled with the debt problems of AMD and in the long run it may be this debt that proves to be AMD’s biggest problem, not scale or innovation.

  3. @ tyson and YJK

    The issue is not AMD’s ability to get close – but how long can they competitive. Intel is still using their money and process lead as a market advantage and AMD somehow just can’t catch up.

  4. AMD is closer behind than it was last time. It took AMD nearly 1 year later than Intel to get to 65nm, if it’s only six months behind then this is a record turnaround time. Besides Intel pays a huge upfront price to get to its next node faster than AMD; AMD will usually get the equipment a little later but the equipment will have some additional development done on it that makes it a new revision.

  5. “which leads me to wonder how long AMD can play this game.”

    Very good question to ponder. Following Moore’s Law will still be very important for a while, but it’s getting incredibly expensive for semiconductor manufacturers to upgrade their equipment. AMD may be relegated to producing specialty chips in the future, not CPUs that stand to gain the most from smaller processes.