More clarity on capex outlook (from CSFB)
Company maintaining C03 capex, C04 looking flat. At the INTC Spring Analyst Meeting last week, the company indicated it would slow its planned 300mm ramp from 50% of MPU production to 20% of MPU production by year-end 2003. The takeaway was misperceived as a cut to C03 guidance for $3.5-$3.9bn. We continue to believe capex guidance was unchanged and see a linear trajectory toward the midpoint of that range. The company is still planning to install 300mm fab capacity as planned in C03, but is planning to ramp production on 90nm in H104 rather than on 130nm in H203 to realize better economics of producing the new Prescott at the smaller line width.
Aggressive phase in of 300mm could drive flat C04 capex. The company continues to plan an aggressive phase in of capex, planning to reach 90% of capacity on 300mm by end of C04. By adding efficiencies on 300mm, the company expects to achieve a 26% capex to COGS ratio in 2003 and 2004, down from 35% capex to COGS during the 200mm ramp from 1995-1999. A capital efficiency of 26% capex to COGS implies a flattish $3.5-$3.9bn capex range based on street revenue and GM estimates for C04. Fabs 24 (Ireland) and D1D (Oregon) will be priorities in C03, with 11X (New Mexico) and upgrades to Fabs 12 and 22 (Arizona) slated for C04. In all, the company has 3-4 fabs that can potentially be converted from 200mm to 300mm using the same copy exact philosophy, although management will not rule out building a new fab if financial incentives make the build decision make more sense.
Moore’s Law focus intact. The company plans to continue to use its technological advantage to achieve the combined advantages of higher GHz processor power with added features (integrated graphics, WLAN, hyperthreading). While the price paid for higher performance has compressed, the higher processor power is becoming a prerequisite to sustain market share, while the integrated functionality is increasingly sought to fill capacity and neutralize the long-term industry ASP trend. Although INTC continues to push down the Moore’s Law curve, we continue to believe higher design and development costs at each successive node will limit the number of chip companies progressing down the curve.
2004 – still hard many chip company significantly ramping spending. Recent analyst briefings all suggest 300mm and partnerships are helping lower capital intensity and are driving less need to ramp capital spending. Our bottom’s up for C04 continues to point to 10-15% y/y growth, below street forecasts for 25%+ y/y capex growth. Although bulls could argue that INTC’s higher 45-50% capital intensity from 2000-2002 still needs to take place for the rest of the industry as it builds 300mm capacity, we believe that 300mm expansion is already underway elsewhere. We are currently tracking 18 projects generating SCE bookings in C03 and have seen 300mm represent 30-50% of order books for companies for the past 2 years. We are concerned that the significant capacity advantage of 300mm will allow chip companies to slowly add capacity to meet more muted end market growth rates, dampening the potential for a significant capacity cycle. |
