Fundamental shift needed

The semiconductor industry is not experiencing an economic cycle but a real need to fundamentally shift away from incumbent technologies and business models born out of the 1980s, to newer technologies and business models that better fit today's market needs.

Three industry dynamics are key indicators. First, OEM development costs are swelling as design complexities grow exponentially to meet new market demands. New semiconductor building blocks are not emerging fast enough to keep pace. Multichip solutions are too difficult to architect and ASICs are too expensive.

The second indicator is the demand shifting to Asia. Asian countries, particularly China, have accelerated their infrastructure to all digital, and now have new users ready to purchase products. Since the volumes from Asia could potentially dwarf the explosion we saw in the '80s in the Western world, a higher rate of adoption is likely to mean much higher demand for products.

The third indicator is that Asian system companies are becoming broader market players. Asia market demands maintain a price-first-centric buying behavior that forces semiconductor companies to reconsider traditional business models. Asian companies have amassed enough domain knowledge and system design capability to become real competitive threats across a large number of system markets that use semiconductors. Western companies are at a competitive disadvantage that cannot necessarily be made up through delivery of semiconductor innovation, because of the rising costs of OEM development.

The U.S. semiconductor model needs to undertake fundamental shifts to regain high growth. More-flexible semiconductor technolo-gies that enable OEMs to amortize system development costs more broadly across OEM product lines must become widely adopted. Broader cost amortization can be achieved only by a shift to an all-software system design approach hosted on programmable platforms (that OEMs purchase off the shelf) so that the costly ASIC development can be dramatically reduced while improving time-to-market.

The industry needs to build companies that couple Western innovation with Asian infrastructure-and create hybrid company models that bring together the best of all worlds. Taiwan and India offer mature and cost-effective engineering that now has to be at the core-not on the fringes-of the U.S. operations model.

Finally, the semiconductor industry must focus on new technologies that improve R&D effectiveness and time-to-money. The time from innovative breakthrough to product in the market must be measurable in months instead of years if Western companies are to effectively use innovation as a key differentiator.