INTC; $20.52; C-2-7
Joseph Osha- Merrill Lynch
GAAP EPS (Dec): 2003A $0.85; 2004E $1.10; 2005E $0.98
GAAP P/E (Dec): 2003A 24.1x; 2004E 18.7x; 2005E 20.9x
Event
Intel acknowledged yesterday that it doesn’t intend to offer a 4 gigahertz version of its P4 desktop processor. Intel had previously planned to offer a 4Ghz part in early 2005, but now plans to top out P4 speeds at below 4Ghz prior to transitioning to dual-core processors in 2H05. Intel also reiterated its intention to stay on an improving performance curve by doubling the amount of cache memory on the P4, to 2 megabits.
Analysis
Changes in Intel’s roadmap normally aren’t worth commenting on, but yesterday’s news bears mentioning because of what it implies for Intel’s roadmap later in 2005. To put it plainly, Intel has a significant gap in its product plans for next year, and the company’s issues with scaling P4 speeds are worsening the problem.
When Intel initially cancelled the 65nm Tejas version of P4 and pulled in its dual-core plans earlier this year, eyebrows were raised owing to the speed with which Intel was planning to move to dual-core. Intel may manage to ship dual-core silicon in 2H05, but the issues associated with transitioning the PC hardware and software communities to dual-core are formidable. We didn’t believe then that there would be a compelling value proposition for dual-core in 2H05, and we still don’t, although we do think that Intel’s shift to dual-core is the right move over the long term.
Now, in addition, Intel is having problems sticking with its plans to update the existing P4 product. Intel is already offering a P4 at 3.6GHz, and it now appears that the product will top out in the very near future at 3.8Ghz. The plan is for Intel to fill the gap next year by offering processors with more cache memory, which will improve performance. Performance doesn’t come free, though, and adding memory to the processor will increase die size, reduce manufacturing yields and put pressure on margins. The approach may work for high-priced server processors, but it is singularly unsuited for the much more competitive desktop processor market.
If Intel runs out of gas during 1H05 on P4, with dual-core processors still 1 to 2 quarters away from a commercially viable ramp, the consequences for Intel’s desktop business will not be good. We have no doubt that Intel has a few tricks up its sleeve to keep P4 performance improving slightly, but investors need to start taking a hard look at Intel’s 2005 product gap. We reiterate our Neutral (C-2-7) rating on Intel’s stock.
