Penn puzzled as chip ASPs fall during boom

7/5/2010 2:03 PM EDT

Perhaps there are other forces at work here, such as doubt. Yes, on the surface, it seems that we're back into boom time, but dig a little deeper and I think you'll find a lot of fear and uncertainty remaining. In the OEM and EMS industries, I still see a lot nervousness. If purchasers at the OEMs are ordering, but still very skittish, than perhaps the chip companies sense that and are using price as a little additional insurance against order cancellations. The thought being, whether well intentioned or not, that that extra bit of loyalty gained by price concessions may be enough to cause the customer to cancel someone else's order. The nervous OEMs are likely using the price card very heavily against the chip companies in order to hedge against their own uncertainty.

7/6/2010 3:33 AM EDT

The interesting item to note is that if higher integration is occurring, with larger SOCs, the inclination is to think the ASPs should be higher on the added cost alone, not counting the profit margins and price aggressiveness. I am not sure how the data is collected and what group of ICs are included (memories, processors, SOCs, analog RF, clocks, and so on). Does a TSV 3D chip count as one (one package), or two (processor + memory) chips. If the latter, then 2 chips with a lower ASP could in fact have higher ASP from they used to be. So the important number is ASP [times] Number of chips.In a consumer driven chip industry, it is all about pricing low enough for the impulse buy. The integrators could potentially get subsidies from their customers for end consumer subscriptions (like say ATT to Apple for phone plans) to keep the hardware affordable, but the bounty does not roll downhill to the IC providers (I think).

I do agree with Duane that sentiment/insurance could be a factor. One other interesting item to consider is that 'silicon is becoming free' when the vendors are bundling services, or sub-systems. So it is not just the chips, but the salsa and the margaritas that come with it that will hopefully sustain the booming business.

7/7/2010 4:45 AM EDT

@CamilleK...clearly you are right about the most important parameter being the value of the market = units x ASPs...and clearly that is going up.....but it is worthwhile understanding that units are growing fast while (some) ASPs are still falling.

Usually ASPs fall on individual chips but the mix changes as new chips are introduced for which a premium can be charged and which overall bring the ASP back up to parity. That used to be a steady equilibrium, then went into a cyclic mode with a period of falling ASPs and then a period of rising ASPs. But for the last several years that has not been happening, ASPs have just fallen and dragged down the total market statistics with them.

In my opinion that is because of the consumerization of the electronics industry over the same period.

One example is that when the microprocessor was at the heart of an expensive enterprise computer it could be priced at several hundered dollars per unit. The 386 went down in price but was replaced by the 486 at a higher price, and so on.

Now the processor is in a consumer product it can only be priced at tens of dollars. So it would seem that reducing ASPs are necessary for, and implied by, that consumerization. But that doesn't mean Penn's concerns for the semiconductor manufacturing industry are not valid.

We just need to realize that there is not a static equilibrium here, it is a dynamic equilibrium of reducing numbers of vendors and there are different time constants associated with different elements of that maturation process.

7/7/2010 6:58 AM EDT

Usually ASPs fall on individual chips but the overall industry mix changes as new chips are introduced for which a premium can be charged and which overall bring the ASP back up. That used to be a steady equilibrium, then went into a cyclic mode with a period of falling ASPs and then a period of rising ASPs. But for the last several years that has not been happening, ASPs have just fallen and dragged down the total market statistics with them.

In my opinion that is because of the consumerization of the electronics industry over the same period. One example of that is that when the microprocessor was at the heart of an expensive enterprise computer it could be priced at several hundered dollars per unit. The 386 went down in price but was replaced by the 486 at a higher price, and so on. Now the processor is in a consumer product it can only be priced at tens of dollars.

So it would seem that reducing ASPs are necessary for, and implied by, that consumerization. But that doesn't mean Penn's concerns for the semiconductor manufacturing industry are not valid. We just need to realize that there is not a static equilibrium here, it is a dynamic equilibrium of reducing numbers of vendors and there are different time constants associated with different elements of that maturation process.

7/10/2010 7:09 PM EDT

It doesn't take a lot to 'puzzle' Penn. He made a prediction and now that it seems to be sputtering...he's confused and not happy. He forgets that times everywhere are getting bad. He believes the United States is not the only show in town and the rest of the world can make up the difference. He forgets or is not informed of the fact that the rest of the world follows what the US does, by and by. So Penn will have to eat some crow before his "predictions" come to fruition.