Seeking Alpha
Profile| Send Message|
( followers)  
Our regular readers know that we have our fears about the state of the semiconductor industry. Since the cliche says that Wall Street is driven by fear and greed, it seems only fitting that our concerns were answered by Andy So of Raw Greed who says:

I once worked for the CIO of GMAM and I fondly remember him repeating his views on technology:

“Technology is a tool and nothing more. Technology is an enabler. Technology in and of itself is not a business. Technology cannot succeed without needed applications.”

The required business formula seems to be Technology + Applications = Probable Success.

We couldn’t agree more with the quote, but when we read it we think more of Pip Coburn’s The Change Function: Why Some Technologies Take Off and Others Crash and Burn. Pip says:

However, a key problem – for starters – is that most potential users of technology products are quite afraid of new technologies. The creators are not afraid, but the users are, and it is the users who matter. SO it seems obvious to creators that this new thing will sell, and then, it doesn’t. Even during the fabulous 90s when nine or 10 technologies finally hit a major commercial inflection, most of the emerging technologies failed to get anything resembling lift off. We have written extensively on 10 failures of the last decade and in no case was technology the problem. All the technologies worked. The engineers indeed developed a cool massive disruptive so-called 10x change in technological capability.

So what went wrong?

Well… we think the implicit thinking – build cool technologies and watch price reduction contribute to generate lift-off for the market – is limited. Creating cool technology is a necessary condition but not sufficient. To consider what is “sufficient” we must bring the user into the equation.

What The Change Framework says is that we only really need to care about what goes on with the users prior to their handing over money for this new thing. Technology demands a change in habits. So… we want to consider why someone would change habits – something I have observed as veeerrrryyyy difficult but that the technology industry implicitly considers rather easy to pull-off.

So… when does a person change a habit?

When the pain from being in a certain state today is greater than the total perceived pain of adopting a solution to today’s pain. If today’s crisis is greater than the total perceived pain of adoption change will occur.

Change = f (user’s crisis v user’s total perceived pain of adoption)

So… let’s suppose [Pip’s nephew] Dylan and my 78-year-old Uncle Jerry have precisely the same desire to have their music with them at all times. Then let’s suppose I offer to help them set up an iPod. I first tell them to go get all their CDs cuz they will need to load them on their computer. Dylan scurries off to get his vast collection. Uncle Jerry says no thanks. He is far more terrified of interfacing with a computer than Dylan. Dylan adopts. Uncle Jerry doesn’t. Uncle Jerry’s total perceived pain of adoption is higher.

In most cases, incidentally, we see price as being less than 10% of the total perceived pain of adoption although folks in the technology world have it burned into their brains that if it is a good idea, all the hold up will be about cost. But, we have been reminded that there is no low enough price for an ugly shirt.

Andy So’s equation leaves users out of the picture, which doesn’t seem to us to be a really logical way of anticipating technology adoption. So what does he suggest will drive technology growth?

We can look at what lead up to the most recent inventory glut in the semiconductor industry and extrapolate which applications are aging. Standard definition DVD, standard definition TV’s, 2G mobile phones, portable audio players and Microsoft Corporation’s (NASDAQ:MSFT) Windows XP associated computer hardware lead the inventory glut in consumer applications. In the business world we saw networking, security, Windows 2000 and Windows XP server and workstation platforms as the leading applications. New semiconductor processes and technologies were created for an aging consumer and business application cycle. These aging applications are experiencing ongoing declining profit margins and demand. Currently we are on the verge of a new application cycle that will undoubtedly refresh both profit margins and demand.

There is very little problem with demand, and we apologize if we ever left anyone with that notion. Demand for semiconductors has been growing at just under 6% annually for the last ten years. This implies far greater than 6% unit growth due to the inevitable price declines technology brings. Our problem is that the 6% growth in demand is being met by 70% growth in supply.

Furthermore, look at the list of aging applications and compare it to the replacements: more of the same. Standard def to high def is an incremental rather than a revolutionary change. So is 2G to 3G mobile or Windows Vista. None of these are the kind of “killer app” that the original spreadsheets and DVD players or the Internet were. Will people buy them? Sure. Will they change the world and usher in a new era of technology wonderfulness? Unlikely. (Note: we have said that if anything will drive a new tech cycle it is Vista.)

Consider, for example, the move to high-def. There is no doubt that the world loves its TV and DVDs. There are also many shows that many people would like to see in high-def. So when it is time to buy a new TV and DVD player, most people will probably buy high-def. But so far, it appears to be just that - replacement demand. By the time the standards war is over, people may have skipped DVDs altogether in favor of downloaded content.

As to Vista, Andy says:

If we take a look at enthusiast websites like HardOCP that have posted a preliminary performance preview of Intel’s (NASDAQ:INTC) Core 2 Quad processor we can already see tangible benefits from running the CPU on the aging Windows XP platform. Multicore processors will be ubiquitous, since there will no longer be any availability of single core processors in due time. The marriage of an OS that natively supports multicore processors with applications that also take benefit of multicore processors represents a staggering cost-benefit relationship that I believe most consumers and businesses will find hard to ignore.

Back to Pip:

Well… we think the implicit thinking – build cool technologies and watch price reduction contribute to generate lift-off for the market – is limited. Creating cool technology is a necessary condition but not sufficient. To consider what is “sufficient” we must bring the user into the equation.

Will the advantage of multi-core processors running code natively be enough of a crisis for most business spreadsheet users to throw away their existing systems and install all new ones, train their employees on how to use them and suffer the lost productivity while the transition is going on? Or will it be something that gets rolled out as the old machines wear out and need to be replaced anyway?

Unlike Andy and his fellow enthusiasts, there just isn’t that much of a crisis to “simply download the free public beta of Windows Vista RC1 here and download a 1080p Windows HD video here.”

And that is why while we agree with Andy’s assertion that semiconductor expansion is still needed, we continue to believe that any capacity growth greater than demand (remember - that is 6% annually) will simply destroy pricing and margins.

SMH 1-yr chart:

SMH 1-yr chart

Disclosure: Author is short put options on DELL and long put options on the Semiconductor HOLDRs (NYSEARCA:SMH).

Comment on this article

Source: Semi Capacity Growth Needs To Slow Down