6 Myths of Semiconductor Manufacturing

Research firm Future Horizons held its annual review and forecast for the semiconductor industry on Tuesday, Dec. 14, in London and predicted the market will grow 6 percent or more in 2011, building on a 34 percent expansion in 2010.

However, Malcolm Penn, chairman and CEO of the UK-based company, was more concerned about certain trends he has identified in the industry, all of which he said have the potential to hamper future growth and create massive supply chain disruptions. Penn said IC vendors are too dependent upon foundries for wafers and warned that capacity is tightening due to sharp cutbacks in semiconductor equipment capital expenditure over recent years.

The researcher identified several myths he says are driving business management in the semiconductor market. These beliefs, according to Penn, formed the background and foundation for decisions industry executives, investors, and financial analysts have taken over the last 10 or more years.

Many of these decisions — including the proliferation of production outsourcing to lower-cost regions of the globe, the use of severe price cuts to win market share, and the fabless manufacturing system adopted by even the largest chip vendors — have upset the traditional equilibrium in the market and further reduced visibility in the already foggy forecasting environment to no more than one to two quarters, according to Penn.

Is Penn right when he says, for instance, that “the single biggest problem we have now in the electronic industry is the huge lack of visibility” into demand at all levels? Electronic companies he added, including OEMs, which used to preside atop the supply chain, now “have no control over inventory while component makers don't know who is buying what because of the presence of contract manufacturers.”

Some of the factors identified by Penn and many other executives and government officials are emerging as major discussion topics across the industry. Officials at the European Commission, for example, are so concerned about the transfer of high-level semiconductor manufacturing to Asia that they recently ordered a review of the benefits and feasibility of reigniting advanced or next-generation chip production in their own region.

That research, lyrically entitled “Benefits and Measures to Set Up 450mm Semiconductor Prototyping and to Keep Semiconductor Manufacturing in Europe,” is set to commence sometime in mid-January 2011. The EC has reportedly awarded the contract for the report.

It's not certain that actions like this can stem the tide of high-tech manufacturing outsourcing to places like China, but people like Future Horizon's Penn, a veteran with almost 50 years experience in the industry, believe very strongly that the West must at least conduct a review into the current and long-term effects of production relocation on its competitiveness.

If you ask Penn to give his opinion, it's clear the facts would have to be undeniably strong to convince him giving up manufacturing by western companies was a smart move. “No one seems to have made the connection that if you give up manufacturing, you set up your economy for deficit financing,” he said. “You cannot grow your economy only on services: You have to do some manufacturing. You don't make money out of research, either. It's good for the brain but not for the pocket.”

Here are the six myths identified by Penn and his rather terse commentary on each one. The myths are in bold followed by Penn's comments in italics. I will explore these individually in future articles, but I would like to know what you think of them. Please post your comment in response to this column.

  1. Fabs have no strategic value.
  2. “Until you haven't got one and you have no control over your business.”

  3. Average selling prices will keep on falling.
  4. “Just like house prices kept on rising?”

  5. Foundry wafer prices will be cheap and freely available.
  6. “Just like cheap debt? Whoops, TSMC is already raising prices.”

  7. Multiple foundry sources will keep the foundry business honest.
  8. “Multi-sourcing at less than 22nm is going to be interesting.”

  9. It's OK to focus on your core business more than on Moore competence.
  10. “Today's more Moore is tomorrow's more than Moore.”

  11. The semiconductor market has matured, and [long-term] industry growth will average 7 percent per annum.
  12. This is the natural outcome of 4 percent falling average selling prices and 11 percent unit growth. However, prices cannot keep falling forever and have been rising over the last six or seven quarters.

7 comments on “6 Myths of Semiconductor Manufacturing

  1. SP
    December 15, 2010

    I agree mostly with the myths listed in the article. Many companies think that investing in fab is not a strategic decision. Its always easy to outsource. Because manufacturing brings lots of worker issues and other production issues. I guess this business needs much higher patience. It doesnt give you much excitement that research gives. And reserach wont give you money immediately. No one can deny the truth that its very important to keep manufacturing control in your hand because ultimately thats the area that brings ideas to reality. So whoever has fabs would dictate the market.

  2. DataCrunch
    December 15, 2010

    I tend to agree with most of Penn’s views and his take on the myths seem to make common sense.  The only one that I am not really sure about is Myth #2: Average selling prices will keep on falling. “Just like house prices kept on rising?” What is the rationale on this one?  Is that prices will not fall because there will be more demand?  But with more demand there will be more production, which means cheaper overall manufacturing runs.  Or is it that as the economies of the countries that Western countries outsourced manufacturing to will significantly improve, which will increase the cost of living and hence, cause prices eventually to keep increasing? 

  3. t.alex
    December 15, 2010

    In fact in the past few years, a number of companies have gone fabless. And typically japanese companies are still retaining their fabs. The trend is more and more startups (in IC design business) are pure fabless and they will go to asia for producing their chips. 




  4. Barbara Jorgensen
    December 16, 2010

    Getting the semiconductor industry to change–or at least question–its business model is a really good concept. But I'm worried it will be like trying to turn the Titanic around. On one hand, the chip industry has a track record of changing with the market. On the other hand, is there an industry–maybe aside from the financial industry–that can be as arrogant at times? I'm not sure if drawing a parallel to the auto industry is even fair, but it's worth thinking about.

  5. pdxsemi
    December 16, 2010

    On “forever falling prices”, I'll say one thing:  in semiconductors, we do this to ourselves.  There is no logic behind creating products with increasing performance, increasing functionality and decreasing footprint for an electronics industry that values increasing performance, increasing functionality and decreasing footprint in its semiconductor components.  As was once said, “in the ocean there's nothing below whale excrement”, and in the electronics industry foodchain, semiconductors are on the same level as whale excrement.  We sit at the bottom of the electronics foodchain, absorbing all the excrement that falls down to us from the OEMs and ODMs.  OEMs and ODMs need price reductions, they turn to semiconductor vendors for fundamental innovation.  They need footprint reductions, they turn to semiconductors.

    Marketing is the reason we're willing to do this.  As a marketing exec in the semiconductor industry, I can say that the worst marketing people on the planet are in the semiconductor industry.  Competitive analysis is usually shoddy at best; definition of product differentiation is, when it exists at all, thin.  Without product differentiation and competitive analysis in which they can have faith, sales teams turn to price to win business.  And so the wheel turns, each loser of market share turns to price to win back that share, dragging down the entire sector of semiconductor vendors with them.  Winning business at lower margins is better than an empty fab.  As I said, we do this to ourselves.

    When everyone assumed that unutilized fab capacity drove price reductions in the semiconductor industry, the fabless model was seen as the business model savior for the sector.  But alas, not having a fab to feed doesn't have any relation to the lack of quality in marketing.

  6. Taimoor Zubar
    December 19, 2010

    @Barbara: I agree that tha auto industry has stood the test of times and has been fairly stable. The growth has been slow but controlled and predictive. I think it's one industry other industries can look up to and may be try to adopt the model.

  7. Himanshugupta
    December 31, 2010

    The most intriguing thought of Penn for me was “You cannot grow your economy only on services: You have to do some manufacturing. You don't make money out of research, either. It's good for the brain but not for the pocket.” I also share the same opinions somehow but i have not analyzed it deeply. Does this thought has any economic backing or just a philosophy? I will love to read a detailed article and findings based on some hard data on the abovementioned quote.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.