It's not a secret that Advanced Micro Devices (NASDAQ:AMD) has had a turbulent couple of years, particularly as its position in the already-weakening PC market has eroded and its competitive position as a second-source to Intel (NASDAQ:INTC) in the server market has been reduced to little more than a footnote. However, after a management transition several years ago, the company has made some important strides toward a recovery — even if it still has a long way to go.
Cutting the Fat and Tacitly Narrowing Focus
Several years ago, AMD tried to have a comparable offering to its much larger and better-capitalized rival, Intel, in both PCs and servers. This spanned everything from multi-socket high-end servers to the lowest end of the PC market. To some extent, AMD still plays in pretty much all of these markets today, but — particularly in high-end servers — it is clear that AMD is getting as much mileage as it can out of existing designs before apparently exiting those segments.
Today, AMD's chief focus seems to be on its "small-core" processors, which have proven useful in a variety of markets, spanning from game consoles such as the Xbox One and the PlayStation 4 to low-power tablet/convertible oriented chips. Of course, big-cores are useful in the more premium segment of the PC market, but they are more expensive to build, and Intel's investment there (and manufacturing lead) is pretty hard to match.
AMD's Latest Low-Power SoCs Look Compelling
While last year's Kabini/Temash are still the latest low-power, low-cost products from AMD available in systems on the market today, AMD did announce a refresh of these products known as Beema/Mullins. While the micro-architecture of these chips seems fundamentally unchanged, AMD has put quite a lot of work into more aggressively pushing the "turbo" capabilities of these chips, allowing Mullins — the tablet and 2-in-1 successor to Temash — to post some pretty impressive benchmark numbers while in AMD's 11.6" reference design.
Now, keep in mind, AMD didn't offer reviewers any measured power consumption numbers, or even any way to measure battery life, so it's too early to draw any meaningful competitive conclusions from the performance numbers alone. In fact, while the power draw of the CPU core at the higher performance levels is probably higher (can't break the laws of physics), the system is likely managing those frequencies better, ultimately delivering a significantly improved user experience over AMD's prior tablet-oriented efforts. Progress matters, and AMD is making it.
We Now Need to See Some Design Wins
Last year's Kabini product was a solid low-power chip for budget notebooks (although Temash definitely needed some work), but the problem was that AMD still managed to lose market share despite these offerings. That being said, while Kabini/Temash were largely uncompetitive against Intel's Bay Trail-M/D and Bay Trail-T, respectively, as far as performance/power goes, Beema/Mullins look like a solid improvement. At the same time, Intel next "big" refresh of its low-power, low-cost oriented tablet and PC products — known as Cherry Trail-T for tablets and Braswell for PCs — won't appear until 2015, giving AMD a window of opportunity.
Of course, it's not safe to assume that because AMD could potentially have the better product (power consumption, delivered performance in real products, bill of materials cost, etc., are all factors outside of the benchmarks that go into this) that it will automatically win designs. However, if what AMD has on its hands is legitimate, and if it can offer these parts at competitive prices to the OEMs, then AMD may have a slight chance of peforming better in the future.
AMD's decision to slim down and focus on a narrower set of products is looking like the right one. While AMD's turnaround is not on track, this is certainly a step in the right direction. However, given the risk/reward ratio, I still would recommend investors to stay away from AMD as the company still has a long way to go before it becomes a good investment.