Merrill out on Advanced Micro Devices (NYSE:AMD) following visiting with AMD and a variety of related PC supply chain companies in both Austin and Taiwan over the last two weeks. The usual near-term chatter about a seasonal uptick in PC shipments has been predictable enough, but a number of more interesting points have emerged from their conversations as well.
Firm continues to be neutral on the stock - there's share gain potential for the long term, but also several intermediate-term challenges that they don't think the market fully appreciates. With that in mind firm offers four predictions:
Prediction 1: AMD's share gain in server processors will continue unabated, with Intel's Woodcrest/Blackford product having only a minimal competitive impact. Firm thinks AMD could get close to 40% of the server processor business by the end of next year.
Prediction 2: AMD won't get past 15% of the mobile processor market in 2007. Firm's meetings indicate that AMD's optimized mobile architecture isn't coming until the fourth quarter of next year, which is too late to have much of an impact on 2007 numbers. The updated 65nm version of the existing Turion product is coming sooner, but they don't think that will have much of an impact.
Prediction 3: Dell will give AMD less business than the bulls hope for the near term, but more than even the biggest bulls hope for over the long term. Firm thinks AMD will see 1.5 million to 2 million units at Dell this year, which isn't much, but they think that AMD should annual volume of 14 to 15 million units at Dell by 2008.
Prediction 4: AMD will allow ATI's discrete GPU business to slowly die on the vine, although management will never be willing to admit it. Firm thinks the 60% of ATI's business that is discrete GPU will show negative revenue growth for the next several years.
ThinkEquity out on Intel (NASDAQ:INTC), saying that they believe the price war may be over. They have heard from several well-placed sources in Asia that Intel has slowed the magnitude of its pricing cuts. This is largely unexpected in the channel. Much of this is occurring through more bundling of higher-end units with lower-end sales. This is radically different that what investors believe, in firm's opinion.
Intel still plans two pricing cuts-October and January. January will likely see Pentium D pricing down 15%-18%. Pentium 4 should see cuts of 10%-20% in January, and Napa declines of 20% (not Napa Refresh). They appear to be holding Merom prices stable beyond January.
Intel is essentially coming to terms with its share loss to AMD. If AMD was gaining share on pricing alone, Intel may be able to win back some of this share with pricing cuts. Firm has heard Intel has come to the conclusion that it will not gain share on pricing, so it should stop cutting pricing as fast as it has.
Intel pricing cuts are more likely at the low end, due to pricing elasticity. It appears that Intel is cutting prices more on the low end, where it really is not competing with AMD. However, it is still planning to lower prices, so the question is: Why? Several sources close to the decision-making process have told ThinkEquity that Intel is primarily interested in driving increased pricing elasticity, rather than competitive pricing cuts.
Firm has heard AMD is in the driver's seat with respect to ASPs. They have heard that AMD has orders above capacity, and it can cherry-pick customers according to margins. While they believe this is not immediately intuitively good for Intel, it removed pressure to cut (and the reward from cutting) prices.
Notablecalls: Two interesting notes on the Intel/AMD situation. Not actionable but good to know category.