DUBAI—The IC industry will shrink from 450 companies to at most 50, with two to three large specialists in each application domain, according to Theo Claasen, executive vice president of business development at NXP.
Speaking to TechOnline after his presentation at the 17th. International Electronics Forum here, Claasen also elaborated upon his comments on industry consolidation, differentiation in the face of common foundry processes and how to achieve maximum profitability. All were hot issues at the IEF closing panel later that morning (see: IC Industry's catharsis in the desert: A mirage gone bad). He also pointed to human interface design and adaptive systems as avenues of innovation that excite him and which he'd like to see explored more.
"We're on the verge of consolidation that has been delayedand I don't understand why it didn't happen before," he said during his IEF presentation. "It's been delayed too long." While this consolidation has started in the U.S. and Europe, "Japan and Korea have defied the trend," Claasen, but he expects that to change. Soon.
For smaller, specialist companies, the driving force behind consolidation is Moore's Law. "For example, codecs are being integrated with functions they [the specialists] don't have, and GPS is being integrated with basebands on phones," he said. "These [specialist] companies will be absorbed," he said. In the end, there'll be two or three large companies in each domain. "And scale matters," he said. "Number 1 and number 2 can make money—number 5 can't." Also, "higher market share equals good profitability," he added, giving the NXP purchase of Conexant's set-top box division as an example.
The merger of NXP's wireless division with that of STMicroelectronics is another. For many years, IC companies have been developing disparate product lines while trying to find synergies between them to lower development costs, but Claasen closed the door on that. "There's no synergy between different application areas," a conclusion that makes even more sense of the NXP/STMicro wireless merger. In general, NXP will consolidate primarily around wireless, consumer and automotive applications, he said, including technologies such as FlexRay and NFC.
Differentiation despite process homogenization
While NXP was all in favor of advanced process technology as a differentiator in the late nineties, when it joined the Crolles 2 alliance, Claasen said it soon came to realize that was the wrong approach. "We had the [advanced process] technology before TSMC, but by 2007 TSMC had ramped it up before us [Crolles 2]," he said. What he realized was that the speed at which a process technology goes from definition to commercialization depends on how many wafers have processed.
TSMC, as a specialist, had processed a lot more and so had rapidly gone through the iterative steps required to get from 2.5 to 0.3 defects per centimeter square. "TSMC did it in half a year; Crolles 2 took one and a half years," said Claasen. Checkmate.
Claasen believes NXP's differentiation now comes through its "system know how," looking at software and hardware in a more holistic system-wide approach. That being said, NXP's expertise in areas such as non-volatile memory can still give it a special process advantage in its partnership with TSMC. However, that special process still costs more if kept proprietary, so NXP has agreed to open it up to other TSMC customers to avoid that penalty, he said, but only after a delay of one year to give it a time-to-market advantage.
Looking forward, Claasen sees ubiquitous wireless connectivity, particularly within the home, as a major technological goal. "We need something more practical than Wi-Fi," he said. "Something that intuitively recognizes the need for a connection." To that end, he sees USB over ultrawideband as the solution.
That need for more intuitive systems carries over to what excites him the most: adaptive systems for human interface design that learn user behavior patterns and help users take fuller advantage of a system's capabilities. "I use 10% of a system's functionality and I've been in technology a long time," he said. "They need to be simpler, more intuitive and comprehensible."
