Critical Undersupply in Semiconductors Coming

So sayeth the head of the World Semiconductor Council.

The head of the World Semiconductor Council warned yesterday that recent aggressive capacity cuts in the D-Ram memory chip sector were likely to result in a severe shortage by the end of the year that could disrupt shipments of computers and other electronics.

The global D-Ram sector faced its biggest crisis in its 16 years of existence last year as industry over-expansion coincided with a drop in demand for computers.

But Frank Huang, chairman of the World Semiconductor Council and also of Powerchip, one of the biggest Taiwan D-Ram makers, said the industry would begin to recover by the second half of this year, and downstream producers of PCs and other electronic products would face "a shortage bigger than I have ever seen in the past 16 years . . . this will affect global PC shipments".

Last year's crisis, reflected in the price of D-Ram falling by more than half to below cost for most producers by the end of the year, led to a store of 1bn unsold D-Ram chips and nearly $10bn in total losses by the industry.

This prompted the biggest manufacturers, mainly in South Korea and Taiwan, to cut production. Others, such as Germany's Qimonda, fell into bankruptcy.

Nearly half of world production had been cut since the beginning of the year, said Mr Huang. This means that by June, the 1bn stockpiled chips will all be sold and by the end of the year there will be a global shortage of 2bn chips, equivalent to the D-Ram needed for 100m computers.

"It has been a difficult time but we can finally see spring coming," said Mr Huang.

Part of the reason for the impending shortage, he said, was that most D-Ram companies were facing financial difficulties and would find it hard to revive production quickly. The other reason was that "wolf has been cried too many times [on a market recovery]. Everyone is too cautious," he said.

I am very, very long semiconductor and semiconductor equipment companies.  They are, by far, my largest industry exposure.  I own a basket of 17 names, including four I have mentioned as net cash plays.

I began building a position in the chips too early, scaled back a bit, started adding more, and have been buying heavily over the past several months.

I have tended to do best buying companies in industries that have been bombed out.   Valuations are extremely low in the semis and some names in my basket could be 5-10 baggers. 

Because of extremely low valuations, I care not what happens in the near-term. I am looking 2-4 years out, not 2-4 weeks or 2-4 months.  So I sit and wait.

Comments

[psychodave]

I was concerned when I read Mr. Huang's
"This means that by June, the 1bn stockpiled chips will all be sold"
for fear this industry representative was just another pump monkey.

With (all figures are real GDP growth from ML, q/q & annualized) 4Q08 posting 6.3% decline, 1Q09 on track for 7.2% decline and 2Q09 estimated to be a 4.8% decline, I feared Mr. Huang's forecast for D-ram demand might be a touch optimistic.

However, your
"I care not what happens in the near-term. I am looking 2-4 years out"
looks like solid thinking to me, and spoke directly to my near term concerns.

[Magellan]

Toro,
I have found your blog my chance a few weeks ago and have been reading since then.
I share a lot of your opinions (although I am not a professional investor and do not work in the finance industry).
As per your post today, sounds like I am not the only one buying semis in the last few months and having semis as the largest industry exposure of the portfolio.
Good to know.
Semis are as cheap as they have ever been and competition in this market is evaporating, as VC are not funding new chip companies anymore and big companies like Qimonda are going bust.
I fully agree that current period of underinvestment (semi equipment book to bill ration at record low of 0.48 in Feb09) will imply higher semi device prices in the future. Add this to the fact that semi market leaders have high gross margins, produce positive cash flows and have large amounts of net cash, semi valuations could soon turn much higher.
I compute a usefull semi valuation monthly chart consisting in dividing the SOX index by the 12-month trailing semi revenues worldwide as per the SIA.
Important data points are as follows:
Low of 1.1 in Jul06
Next low of 1.5 in Aug08
Absolute top of 7.8 in Feb00
Next low of 1.6 in Oct02
A top at 3.4 in Nov03; from there it is all down hill down to low of 0.63 in Nov08.

So in Nov08 the SOX low is 2.5 cheaper than the SOX low in Oct02.
The low in Nov08 is 1.75 cheaper than the previous absolute low in Jul06.
Currently we stand at around 0.92, which is still 1.7 cheaper than the Oct02 SOX low.

[Toro]

Magellan

That is an interesting way of looking at the group. I go on a name by name basis and compare where they have traded in the past primarily relative to book value.

As for the book to bill ratio, one of the Wall Street firms estimates that current capex is running at 40% of MAINTENANCE capex. The chip equipment group is also cheaper than the semis. Thus, semi equipment makes up the majority of my basket.

T.