AMD announced their Q1 2012 earnings last week, which turned out better than the previous numbers suggested. The bad news is that they posted a net loss of $590 million. That does sound pretty bad considering that their gross revenue was $1.59 billion, but there is more to the story than meets the eye. Of course, there are thoughts of “those spendthrift executives are burying AMD again”, but this is not the case. The loss lays squarely on the GLOBALFOUNDRIES equity and wafer agreements that have totally been retooled.
To get a good idea of where AMD stands in Q1, and for the rest of this year, we need to see how all these numbers actually get sorted out. Gross revenue is down 6% from the quarter before, which is expected due to seasonal pressures. This is right in line with Intel’s seasonal downturn, and in ways AMD was affected slightly less than their larger competitor. They are down around 2% from last year’s quarter, and part of that can be attributed to the continuing hard drive shortage that continued to affect the previous quarter.
The biggest news of the quarter was that AMD is no longer constrained by 32 nm availability. GLOBALFOUNDRIES was able to produce as many 32 nm parts for AMD as needed with yields continuously improving over the past two quarters. AMD seems very comfortable about where they are at in terms of yields and availability for both Bulldozer and Llano based product lines. AMD has in fact been ramping production of the upcoming Trinity based processor and has been shipping finished products to customers since mid Q1. They have also started shipping Brazos 2.0 parts to customers, and both Trinity and Brazos will be launched in mid Q2 of this year.
The CPU/APU World According to AMD
The mobile area has been one of tremendous growth for AMD and Q1 saw 100% of all mobile shipments be APU products (both Llano and Brazos 1.0). AMD is very bullish about Trinity. They say that it offers around 50% more performance at the same TDP as the earlier Llano based processors. This 50% is a combination of both CPU and GPU performance, so do not expect massive jumps in CPU performance alone from current Llano based products at those TDPs. The big jump does appear to be in graphics, and AMD is certainly more than willing to hang their hat on that portion. With the latest Ivy Bridge IGPs still not able to match last year’s Llano, AMD feels that Trinity will truly leave Intel behind in terms of overall graphics performance. Trinity features a totally redesigned graphics portion which combines the VLIW4 architecture of the HD 6900 series with aspects of the new 7000 series of products.
Brazos 2.0 will still be based on 40 nm parts, but they apparently have been heavily redesigned to improve performance and power consumption. These new chips will be combined with a new Fusion Controller Hub (FCH) that will add more features to the mix. AMD was sparing with the actual details of this product, but they are really pushing them into the tablet and low end notebook space. For the past year we have seen a surprisingly robust selection of E-350 and E-450 parts in sub-$500 notebooks featuring large screens, big hard drives, and higher memory densities. Brazos 1.0 has been one of AMD’s most successful lines in recent history.
AMD will not be charging obscene amounts of money for their Trinity based products, and they are hoping to really make a dent into what they consider to be the meat of the market; namely the area between $400 and $700 for notebooks. Trinity will offer TDPs down to 17 watts, which will allow it to be integrated into thin and light products, or what AMD refers to “Ultrathins”. This is the same style of chassis for which Intel is spending millions on promoting with their “Ultrabook” branding. AMD is hoping to be able to offer products that are around $200 to $300 less expensive than comparable Intel branded Ultrabooks. If AMD is able to achieve this with a strong enough chassis design, features, and heat dissipation then they will offer a tremendous value as compared to what Intel offers. Trinity will also feature aggressive power saving techniques and will also include the AMD Video Codec Engine, which in theory is competitive with Intel’s Quick Sync technology.
AMD claims that they have seen a record number of design wins for their Trinity and Brazos 2.0 parts. Demand is already very healthy, and we can see why AMD is bullish about these two products. AMD also claims that they have a significant number of design wins for the thin and light category, and we can assume that these will feature both Trinity and Brazos 2.0 APUs. Now, whether this matches the 26 design wins that Intel has so far for their Ultrabook category of parts is unknown. We can expect these designs to be unveiled quite soon though, and by soon I mean before June 1.
The mobile market has been incredibly important for AMD for the past four quarters, and now that they are no longer supply constrained on either 40 nm for Brazos or 32 nm for Llano and Trinity, they really hope to explode into the market. The desktop market, on the other hand, has been a mixed bag. Llano sales have been solid, but Bulldozer has been a pretty big disappointment. While AMD did not mention it in the conference call directly, we look at the numbers and see that the desktop Bulldozer is a costly product to make in terms of die size. AMD still offers 45 nm quad core products to take up much of the slack left by Bulldozer. Llano is also still selling strong on the desktop due to its still class leading integrated graphics performance. There will be no new Piledriver based CPUs for the AM3+ desktop market in Q2, but we can expect them in Q3. We can expect an aggressive push for FM2 and Trinity based desktop parts in June though.
One area that has been overlooked by most has been that of AMD’s server market share. They have seen three consecutive quarters of growth from this area, and a lot of it has been due to the much maligned Bulldozer architecture. While Bulldozer is not fantastic for the desktop, it achieves much better results in many enterprise class applications. While AMD’s product portfolio is nowhere near as strong as Intel’s, it has achieved sales growth that has not been seen since the original Opteron. The Interlagos products, which feature two Bulldozer dies together on one substrate, is a good combination of TDP, performance, and thread count for the single and dual socket market. Four socket Interlagos offers a compute density that is certainly hard to match for the price. AMD will update their server lineup with Piledriver based parts, but that again will not be until a Q3/Q4 timeframe.