INTC: Intel to (Over)Invest $1.5B in N.M. Factory
We are not big fans of excess capacity. When there is too much capacity, there are too many products and that tends to push profits to about zero. We’ve seen it in airlines, we’ve seen it in autos, and we’ve seen it in tech. But most of all we’ve seen it in semiconductors, which is why we hate to see stories like Intel to Invest $1.5B in N.M. Factory: Financial News - Yahoo! Finance:
Intel Corp. plans to spend between $1 billion and $1.5 billion to overhaul its semiconductor production facility in New Mexico to manufacture computer chips with next-generation technology.The Rio Rancho factory is expected to begin producing 45-nanometer chips — meaning they will have features as tiny as 45-billionths of a meter — in the second half of 2008, Intel said Monday. The transistors on such chips are so small that more than 30 million can fit onto the head of a pin.
Shrinking the circuitry of microprocessors, which act as the core calculating engines of computers, is essential to keeping up rapid gains in performance and energy efficiency.
Smaller chips let companies cram more transistors onto each slice of silicon and maintain the pace of Moore’s Law, the 1965 prediction by Intel co-founder Gordon Moore that the number of transistors on a chip should double about every two years.
So they’ll be able to put more transistors on each chip and more chips on each wafer. The big question remains: do they need more chips and/or transistors? If they do, then clearly increasing capacity is the right thing to do. If they don’t then clearly it is the wrong thing to do. To answer the question is fairly simple: we look at inventories. If they are making too many chips inventory levels will go up.
As the chart shows, inventories are going up. The last thing the microprocessor industry needs is more capacity. In particular, Intel doesn’t need any more capacity. We said before that the biggest potential positive from private equity buyouts of semiconductor companies is that they might instill some financial discipline. So we can only help but wonder if the wrong company is in play.
Disclosure: William Trent has a long position in SMH.
Like this article? Why not try out:
PLEASE TELL ME YOUR FORCAST FOR MICROPROCESSORS IN H2 2008, THE TARGET OF INTELS 45 NM UPGRADE IN NEW MEXICO.
THEN EXPLAIN THE RELATION OF THAT DEMAND TO THE CURRENT INVENTORY. 18 MONTHS IS A LIGHT YEAR IN THAT BUSINESS.
My best answer to that is probably this post.
Unit demand is likely to rise high single- to low double-digit each year - so perhaps 20% over the time frame you reference. However, since January 2006, spending on new capacity has been rising far faster than that.
What do today’s inventories have to do with the forecast? Well, last year’s capacity was obviously too much since inventories grew. Nonetheless, last year the industry ordered new capacity at a faster rate than they were able to sell it - so this year there will definitely be too much. And now we see they are already working on their 2008 capacity - just to make sure there is too much of it, I guess.
At this rate, be prepared to wait a light year before investing in semiconductor manufacturers.
hmmm …
on the other side:
If you are the market leader, which Intel is, you need to keep investing. If you stop doing so, you will fall behind.
What Intel is doing, is using economy of scale.
By using 45nm the can increase their margin compared to competition. They are also in an easier position to decrease pricing. At the end of the day, in the current situtation it will keep AMD from decreasing their prices too agressivley because they know Intel can go even deeper, if needed by far.
Should Intel be able to maintain a lead over AMD for a couple of quarters, this will have big effects.
Another perspective is this:
A good CPU (or PC if you will) costs always the same. You just get more for it than 12 month before. So rest assured, that Intel will find something to fill up the siliconscape of a CPU. This means, that even with 45nm the size of the chip may stay similar because Intel has added some other bells and whistles.
For me this all means stormy weather for AMD. Not now, but the bad weather is not too far away.
Intel is working an decoupling themselves from AMD, AMD might need to find a new business modell in 2-3 year from now. Maybe they need to merge with another big company to stay alive. Who knows. I don’t be honest, but I think future will have some suprises for us.
Continuing to invest is one thing. Investing more than is necessary is another, and it is this “other” that Intel (along with the aggregate industry) has been guilty of. When I measure capacity plans I consider them in relation to sales growth. Over the long term they will grow in a similar fashion. If capacity grows faster there is too much supply, and if it grows slower there is not enough. The extent to which there is an imbalance (how long and how large) determines how out-of-balance the picture is.
Excess capacity is bad for all players in the industry (see autos, airlines, et. al.)
I agree, excess capacity is bad for all players.
But please keep in mind, that those CPUs are not just a “regular” product or service like autos or airline seats. It’s high-tech stuff, and here the market can be very dynamic and bend the usual rules.
Another effect of excess capacity is, that a competitor may be left behind crippled because it cannot keep up. Kind of an arms race.
And for a more medium term time range you may find that Intel has put something else on the chip. Right now additional CPU-cores and more cache is “en-vouge”. In future it will be other virtual-realtity type of stuff. The 80 floating-point core research chip from Intel shows the direction. You will not see an 80 core chip from Intel, but you might see regular dual core x86 CPU plus multiple special cores on the die. This way the PC becomes more powerful and prices stay the same (while the size of chip acutally stays the same even if geometries shrink).
Hopefully, also the market, especially Asia will gain more ground. More people move from the country to a (newly built) city. This requires infrastructure, infrastructure this days is coupled with high-tech and IT.
I am not quite what the effect of the upgraded fab will be, but I am quite sure it will improve Intel’s competetive position in the industry.
It all remains to be seen, and of course my opinion is more like a gut feeling from my side since I am not an analyst.
I agree, excess capacity is bad for all players.
But please keep in mind, that those CPUs are not just a “regular” product or service like autos or airline seats. It’s high-tech stuff, and here the market can be very dynamic and bend the usual rules.
Another effect of excess capacity is, that a competitor may be left behind crippled because it cannot keep up. Kind of an arms race.
And for a more medium term time range you may find that Intel has put something else on the chip. Right now additional CPU-cores and more cache is “en-vouge”. In future it will be other virtual-realtity type of stuff. The 80 floating-point core research chip from Intel shows the direction. You will not see an 80 core chip from Intel, but you might see regular dual core x86 CPU plus multiple special cores on the die. This way the PC becomes more powerful and prices stay the same (while the size of chip acutally stays the same even if geometries shrink).
Hopefully, also the market, especially Asia will gain more ground. More people move from the country to a (newly built) city. This requires infrastructure, infrastructure this days is coupled with high-tech and IT.
I am not quite sure what exactly the effect of the upgraded fab will be, but I am quite sure it will improve Intel’s competetive position in the industry.
It all remains to be seen, and of course my opinion is more like a gut feeling from my side since I am not an analyst.
Tech companies only think they are different from autos or paper stocks. If you look at the chart, they are cyclicals, and cyclicals are supply driven.
Thanks for the comments. It’s nice to see that people are reading and thinking about what I’m saying, and goodness knows I’m not always right. If nobody comments I’m just another guy in an ivory tower. Even if I disagree with a comment, I should have to back up my opinion.
Hope you continue to enjoy, and discuss, my articles!