To: fyodor_ who wrote (53189 ) 8/30/2001 11:15:06 AM From: AK2004 Respond to of 275872 09:44am EDT 30-Aug-01 SoundView Technology Corp. (Scott Randall 203-462-7246) Advanced Mirco Devices (AMD) Estimate Changes August 30, 2001 Advanced Micro Devices (AMD) Rating: Buy Price: $14.20 'Fiscal Year Dec. 31 F00A F01E F02E 2Q01E 1Q00A 2Q00A EPS ($) 0.27 0.33 -0.13 Old EPS ($) 2.34 0.41 0.64 -0.07 2.34 0.41 Revenue ($M) 3945 4043 846.6 Old Revenue ($M) 4644 4020 4384 879 4644 4020 In conjunction with SoundView's semiconductor conference yesterday, AMD effectively lowered its guidance to the lower end of its range. While AMD had been guiding to a Q-Q revenue decline of between 10% and 15%, its guidance is now for a 15% decline. AMD noted that Flash is expected to decline by about 30% Q-Q or by $100 million vs. earlier guidance of up to a 30% decline. With AMD noting that its unit shipments of processors are still expected to be at record levels, the effect of the current price war is serving to reduce revenue contribution more than expected. In our model, we have lowered our ASP assumptions for the September quarter to $70 from $74, while leaving our unit assumptions unchanged. With both AMD and INTEL willing to defend their respective market shares almost at any cost, pricing pressure will continue to ratchet through both vendors. With both AMD and INTC having excess capacity, PC OEM's fighting a fierce price war, weak PC demand and with both AMD and INTC shipping competitive products, we expect the current environment to continue. No change in our view of the challenges facing PC component suppliers. AMD's commentary yesterday (and its change in guidance to the low end of the range) reflects the difficult market for PC component vendors. Although we believe that PC component vendors have seen some pick up in demand, we believe that this pick up has been less than would normally be expected from a seasonal perspective. We believe the current aggressive pricing environment for AMD and INTC will continue for the following reasons: * AMD and INTC are shipping competitive products * AMD and INTC have excess capacity * PC OEM's are fighting a fierce price war * PC demand is weak In flash, AMD notes that it has seen some signs of improving demand from wireless component vendors; which is consistent with our earlier comments. AMD's goal is to maintain its microprocessor market share (at about 22%) through the end of the year. As we noted returning from our trip to Taiwan this spring, INTC's aggressive pricing is serving to stem AMD's market share gains. AMD's presentation yesterday focused on AMD's response to INTC's current clock rate superiority at the high end. In order to compete with INTC, AMD spent considerable time discussing its plans to combat Intel's current clock rate advantage. Recall that currently, AMD's highest clock rate shipping product is 1.4 GHz while, Intel has now announced a 2 GHz device. From an engineering perspective, clock rate is but one determinant of processor performance. By both AMD's observation and Intel's admission, the Pentium 4 is in fact a less efficient engine when compared to previous generations of processors. Stated another way, the Pentium 4 does not deliver the performance users would expect by a simple scaling of clock rate. This is due to an architecture, which has sacrificed clock rates for efficiency in term of the number of instructions that can be completed per clock. In comparison, AMD's Athlon has maintained a higher efficiency in the number of instructions that can be executed per clock while contributing to a slightly lower clock speed. Consumers and/or investors should consider AMD's challenge of convincing the market that a slower clock rate device delivers better performance than an Intel higher clock speed device. AMD's relative success at this effort will, we believe, be a critical determinant of the Company's success over the next several quarters. AMD's effective challenge will be to both educate and then convince consumers that there should be more to their processor buying decision than simply MHz. While we believe this is a technically sound argument, it is challenge for AMD to achieve. Recall that several years ago, CYRIX effectively faced the same challenge of establishing a performance rating (PR rating). This PR rating was used to identify a processor by its effective speed rather than its actual clock rate. This approach achieved limited success in the market. Instead of a PR rating approach, we believe that AMD will focus on demonstrating (at the point of sale for consumers) the benefit of real world applications for systems, based on its slower clock rate processors having better underlying performance. Although we believe that this approach has merit, we believe that the challenge for AMD will be significant.