The Stakes Are High

By Joanne Itow
Whether you call it a transition or a transformation, evolutionary or revolutionary, the implementation of 450mm wafers is going to be disruptive. There are both technical and operational hurdles associated with semiconductor production on 450mm wafers. The ramp of 450mm is expected to change the semiconductor manufacturing landscape. And what about the changes that may occur even before 450mm is implemented?

Recently I’ve been asked about the potential for 300mm wafer shortages over the next five years. We’re already hearing that 300mm wafer production capacity at the major suppliers is filling up. What incentive does a silicon wafer supplier have to build additional 300mm wafer capacity? Over the next five years, Semico forecasts total wafer demand will increase at a compound annual growth rate of 9.3%. Demand for 300mm wafers is expected to grow even faster than the overall total. Will the existing capacity be enough to carry the industry through to 2017/2018 and beyond?

Even if a silicon wafer manufacturer wanted to build a new wafer plant, who would finance it? At a recent SEMI Arizona Steering Committee meeting, Dan Tracy, senior director for industry research at SEMI, reported that silicon wafer revenues declined to $9.1 billion in 2012 from $10.4 billion in 2011, a 12% drop. Although silicon unit shipments increased, average selling prices declined even faster. That’s not a good situation for any material supplier. In contrast, organic substrates used in packaging grew 6.1% to $10.4 billion in 2012, compared with the previous year. Things have certainly changed when the value of the packaging materials outpaces that of the basic semiconductor substrate.

It is unlikely that a major silicon supplier would build new wafer capacity in the next five years, except for possibly the first 450mm wafers. If silicon wafer supply does get tight, and if basic economic principles of supply and demand are allowed to operate, prices of 300mm wafers should begin to increase. And higher costs for existing silicon substrates might help justify the economics for 450mm wafers.

But what if 450mm is delayed beyond five years? This increases the probability that another supplier steps in. What if a silicon supplier in China or Korea builds additional 300mm wafer capacity? That would ease the supply situation and keep prices from skyrocketing. Unfortunately, it also would mean the suppliers that are committed to supporting 450mm wafers would suffer from lower margins. The timing of the semiconductor production ramp on 450mm could be devastating to the major wafer manufacturers.

This situation truly exemplifies the fact that the economics and timing of a major industry transition can be much more challenging than the technology transition itself.