Dave,
As usual, thanks for the thoughtful posts.
I don't have nearly as much empirical evidence at my disposal as you do -- I know a fair amount about the technology world generally, but you clearly know more than I do about the SEMs in detail, and that's just one more reason to respect your opinion.
IMHO, there are several potential mistakes, though:
(1) You seem to think demand is unresponsive to price cuts, that dramatic price reductions in DRAM won't really do that much to boost demand or profits, even when coupled with dramatic cost reductions. (E.g., "there are more chips out there than people really need", "they can crank out more chips and nothing changed in the demand equation") Really? I know I buy a lot more RAM (etc.) when it drops in price, given that RAM substantially improves performance. And I don't use voice, but common bloatware like Corel Draw, Office 97, and IE 4. As Internet high-bandwidth connections and voice processing gradually become ubiquitious, we'll need so much more horsepower than we have at this time that it won't be funny. (True, I don't have a detailed timetable, but obviously it's going to pick up a lot every year.) There is no fixed quantity of needed computing power, and just as clearly low prices are one of the biggest ongoing spurs for demand -- always have been, always will be.
(2) You apparently think that rapid growth in computing demand based on high-power uses is also dubious, even though that's been the way things have worked in the past, and even though the rate of software and hardware innovation is faster now than ever before (again: voice, Internet, let's add biometrics, video [I know, not today], telecom [holy cow], medical, 2d/3d world development [SE Asia hurt this, temporarily], and many more that I don't know about about). This seems merely an assertion, unless you are simply talking about the very short term. (E.g., unless they add voice to W98 -- they may -- I doubt many W95 users will upgrade, except to get FAT32.)
(3) You seem to think INTC is the PC industry ("If the [INTC] MPU sales are slowing....than [sic] we are talking about 60% of the chip industry slowing"). But the whole reason for "fear" for INTC's future (i.e., fear that their future is only extremely rather than maximally rosy) is that they're missing out on one of the current hot trends (sub 1k PCs), not at all that the PC industry itself is tanking. The industry has more demand than it can handle in many respects (Cyrix, AMD, and PCs based on them, as two obvious examples), even if INTC doesn't (at their higher price points, anyway; that'll change when they drop prices).
To put it another way, if INTC slows (i.e., their growth slows), it's not that 60% of the market will slow, it's that Intel will lose market share, as they've been doing this year. Lack of consumer demand is not at all directly Intel's problem-- competition is.
Others have challenged the -overall- fab capacity issue (no one contests that at this instant there's too much DRAM wrt its current production cost; your claim of an MPU surplus, or a general chip surplus, seems dubious to me), but those are more strictly empirical issues, so I'll leave those to the folks with easier access to the data.
Sorry that my comments are fairly -qualitative- rather than -quantitative- comments, which means its difficult to draw (short-term) quantitative conclusions from them. I just don't have time to become a SEM economist.
And don't get me wrong -- I think your cautionary notes are good ones. But I think they're a bit overstated, and definitely short-term more than long-term concerns. I'm not one of those who says we'll set new highs in 3 months, so maybe we don't disagree. But I want to make sure we don't temper over-optimism with over-pessimism either. Medium- and long-term, I think we agree that the key to profitability is dramatically increased demand, which I think will almost certainly arise.
I need to get back to my software development work now, for Intel's sake! :^>
--John |
