It was interesting to hear the JPMorgan semiconductor analyst ask ST CEO Carlo Bozotti at the Q3 earnings call whether the fab-lite trend is on the wane.
“Surprisingly, the trend seems to be changing in the semiconductor industry,” the JPMorgan analyst told Bozotti, “many of your other competitor IDMs are talking about building more fabs at this time. So, I mean, what are your thoughts on that, as well as, whether you can use your 300-millimeter capacity at Crolles, given that, I mean, Texas Instruments et cetera keep talking about large 300-millimeter fabs in analogue, whether you need that sort of capacity?”
The JPMorgan man’s question is interesting not because the thoughts of Wall Street on semiconductor industry strategy have ever been particularly insightful, but because the Wall Street analysts hunt as a pack and, for several years, the pack has subscribed to a consensus judgment that fab-lite is the inevitable future of the industry.
If the question from the JPMorgan guy suggests that the pack consensus is changing, that is significant, because the more feeble-minded CEOs pay a lot of attention to the Wall Street pack.
For years the semiconductor industry’s most authoritative analyst, Malcolm Penn, CEO of Future Horizons, has argued that the fab-lite strategy was a disaster waiting to happen.
This year, with the automotive industry, the telecoms industry and the consumer electronics industry all complaining about a shortage of ICs, Penn was proved to be right.
In September 2009 Penn warned: “Lack of investment in fabs will strangle the IDMs to the sole benefit of the foundries. Just as the world has suddenly rediscovered cash is king, it will only take a capacity shortage to reinforce the message capacity is king … that shortage is coming.”
And how right he was!
NXP doesn’t expect any growth this year in a 30% growth year. That’s because it lacks manufacturing capacity after savage cuts and closures at fabs in Caen, Hamburg, New York and Nijmegen in pursuit of a fab-lite strategy.
Last month Linear Tech CTO and co-founder Bob Dobkin ripped into the fabless business model as being detrimental to the quality of ICs.
“We’re one of the few semiconductor companies that make the majority of their chips in their own fabs and have the ability to tailor the process to the circuit, which gives us the best products possible,” said Dobkin, “companies using foundry processes don’t have the same control as we do for things like noise.”
“People think of small line widths as the highest form of technology, but that’s not true,” added Dobkin, “the highest form of technology is in circuit characteristics like stability, precision, power-efficiency and noise – these things make the difference between high quality circuits and those that give you surprises.”
“You really have to keep track of the quality and the performance,” Dobkin continued, “innovation resides either in the circuitry, or in the complexity of the circuitry, or in the combination of the circuitry and the process. You have to combine the circuitry and the process to get the highest quality device. If you don’t have a foundry you have to see how much circuitry you can put in there. To make high quality chips you have marry the process with the design.”
And to make any chips at all at a time of a shortage of capacity you have to have access to a fab.
Brilliant link, it does explain a lot about the craziness of “the markets”.
As for fab-lite, it’s a nice idea to gain an edge on your competitors – but only for a short while. As soon as everyone tries to go fab-lite the idea no longer makes any sense. I guess there are all kinds of examples of things that are good marginal strategies that break down and do damage once they become main stream, and this is just another of the same.
Who started it all anyway? It wasn’t Fred Shlapak, was it?
Thank you George, it has been corrected
David, in the sixth paragraph, you may wish to change “auromotive” to “automotive”.
SemiCurious, thanks for the interesting link. It makes more sense than most of what I have read about Wall Street. It is more logical, too.
“the Wall Street analysts hunt as a pack”
This is what the hormone powered, MBA empowered Wall Street boys, in the end, are really after:
http://krugman.blogs.nytimes.com/2010/10/30/sex-and-drugs-and-markets-role/