Lowering 2005 EPS Estimate to Reflect Revised Roadmap - Morgan Stanley
Although we just lowered our 2005 revenue and earnings estimates when Intel reported its third-quarter results, we are making another cut, as we believe that a change in the company’s product roadmap will likely pressure processor average selling prices (ASPs). Based on this expectation, we have lowered our 2005 EPS estimate to $1.05 from $1.10, and our new estimate compares to the consensus estimate of $1.12. If Intel is unable to introduce and ramp its dual-core processors into volume in a timely fashion, then we believe that additional earnings reductions are
likely to occur. Regardless, we believe that the vulnerability we perceive in Intel’s current product line up will result in an increase in revenue, margin, and earnings risk next year.
Based on the reduction in our earnings estimates, we have lowered our 12-month stock-price target on INTC to $23 from $25. Our target assumes that the stock should be fairly valued at a PEG of 1.5x our earnings and 4x our revenue estimates in 2005. The average stock in our universe is currently trading at 3.0 times our 2005 sales-per-share estimates, and INTC’s 10-year average P/S multiple is 5.5. We maintain our Equal-weight stock rating, and believe that INTC will find solid intermediate-term support in the mid to high teens, as the stock’s 10-year low price-to-tangible book value multiple has been about 3x, and tangible book value, which is currently $5.28 per share.
Pentium 4 has reached the end of the line…
Intel has started to brief its customers on its updated product roadmap, and the company has dropped 4 GHz and higherspeed versions of its Pentium 4 family from its future product plans. Intel introduced its Pentium 4 architecture in 2000 and the current version (known as Prescott) was introduced in the first quarter of this year. At the time of the introduction, Intel expected its Pentium 4 architecture to eventually be able to scale to 10 GHz speeds. However, ever since Intel began to transition its Pentium 4 architecture from Northwood (0.13-micron design) to Prescott (90 nm design), the company has struggled to
introduce the various members of the Prescott family in a timely fashion. We believe that the challenges that have plagued the execution of the Prescott roadmap have been caused by the design’s inefficient performance versus power dissipation characteristics.
Despite our concerns about the future scalability of its Pentium 4 architecture, various members of Intel’s management team assured us that they would continue to introduce higher-speed versions of its Pentium 4s throughout 2005 when we met with them at Intel’s Developer Forum (IDF) in early September. However, the
company’s earlier delay in the introduction of its 4 GHz product and its recent termination suggest that our concerns about the additional headroom of the Prescott design were well founded. Given the premature termination of Intel’s Pentium 4 roadmap, we believe that the company’s overall product roadmap will be more vulnerable than it has been since early 2000 when the Pentium III architecture began to run out of headroom in early 2000. At that time, Intel’s Pentium III was unable to exceed 1 GHz speeds, company struggled until it was able to introduce and ramp up its Pentium 4 processors.
…And the potential for a gap in its MPU product line suggests that ASP pressure should increase in 2005 The high end of Intel’s desktop PC processor product portfolio currently consists of a 3.6 GHz Pentium 4 with 1 MB of cache memory. We expect Intel to introduce a 3.8 GHz Pentium 4 with 2 MB of cache memory in the first quarter, and this should be the highest performance member of the Pentium 4 architecture to be introduced. Thereafter, we believe that a gap between higher performance versions of Intel’s desktop processors will likely exist until the company can introduce its dual-core processors in the second half of next year. Thus far, Intel has been unwilling to provide any detailed information about its dual-core product introduction and ramp-up schedule. While we currently expect Intel to provide detailed information about its dual-core ramp plans at its Developer Forum in February, it is certainly possible that the gap we perceive in the company’s product portfolio will last longer than we
currently expect.
Over the years, Intel has used its ability to introduce higher performance processors to manage its overall ASPs. Following a steady decline in ASPs in the 1997-1999 timeframe, Intel’s processor ASPs stabilized until the end of 2000 and then declined sharply again in the first half of 2001. Since then, we estimate that Intel’s processor ASPs have remained in a pretty tight range of approximately $150-$160. From our perspective, we view Intel’s pricing strategy as pretty straight forward, as the company reduces prices one to two times per quarter. However, these price
declines almost always coincide with the introduction of a higher performance processor that enters into the high end of the company’s well-defined price bands. For example, when the 3.6 GHz Pentium 4 was introduced earlier this year, it initially occupied the $637 price band, while the prior high-end processor (3.4 GHz) was reduced from $637 to $417 and the 3.2 GHz part was reduced at the same time from $417 to $278, and so on.
Intel has successfully used the above strategy to maintain relatively stable ASPs during the last several years. With stable ASPs, Intel’s gross margins primarily fluctuate with capacity utilization rates, die sizes, yields, and overall capital intensity. Since Intel will have largely completed the transition to its 300mm/90nm manufacturing capacity by the end of this year, any changes in ASPs can be expected to have an impact on gross margins in 2005. With increased competition from AMD and the potential that Intel may be unable to refresh the high end of its product line as rapidly as in the past until its dual-core processors are introduced, we believe that the company’s overall ASP will likely decline versus our prior expectation for a relatively flat trend in 2005. We estimate that every $1 decline in Intel’s processor ASPs will result in a 45 basis point reduction in the company’s gross margin and nearly a $0.02 decline in EPS in 2005.
Key risks
The risks to our earnings estimates and investment thesis are predicated on the potential that the current economic trends and overall PC demand may be weaker than forecasted. Since Intel largely operates a fixed-cost business, solid unit demand and favorable ASPs are essential to support attractive gross margins, and INTC has historically proven to be a gross margin sensitive stock. In addition, Intel’s product mix will continue to have a
significant impact on the company’s overall margins, and a sustained shift to lower-end PCs or Intel’s inability to introduce higher performance/higher priced processors in a
timely fashion would likely create further earnings risk. We also note that as device dimensions are scaled to smaller sizes, issues such as leakage current (leading to higher power dissipation) need to be closely monitored and an inability to contain such device degradation could pose an adverse risk to the introduction of succeeding technologies.
