Issue



It’s time for new innovation


2017-06-20 13:41:36

BY PETE SINGER, Editor-in-Chief

What if the automotive industry had achieved the incredible pace of innovation as the semiconductor industry during the last 52 years? A Rolls Royce would cost only $40, go around the world eight times on a gallon of gas, and have a top speed of 2.4 million miles per hour.

That point was made by Subi Kengeri speaking at The ConFab in May. Kengeri is vice president, CMOS Business Unit, at GlobalFoundries. He also noted that if one of today’s high performance graphics chips were produced using 1960 vs now “it would be the size of a football field.”

Clearly, no other industry can match the pace of innovation of the semiconductor industry. “The transistor count per square inch in 1965 was roughly 100. In 52 years, if you follow Moore’s Law of 2 years per innovation cycle, that gives 26 innovation cycles. That’s 100 millionX improvement (2X26),” Kengeri noted.

Of course, there has been plenty of innovation in the automotive industry. Interestingly, most of the exciting new innovations such as backup cameras, collision avoidance, navigation/ infotainment, self-parking, and anti-lock brakes are only possible because of semiconductor technology.

Kengeri said that Moore’s Law scaling will continue – “there’s no question about it,” he said – but there’s a growing need for new innovation to address the increasingly diverse array of semicon- ductor applications. These are driven by growth in mobile computing, development in IoT computing, the emergence of intelligent computing and augmented/virtual reality.

“Leading edge innovation will continue and all the leading manufacturers continue to invest, whether it is litho scaling in terms of EUV, or device archicture,” Kengeri said. “What is really important is how to do we continue to innovate, how do we continue to get the value at competitive costs? Trying to get the scaling at any cost is not what is needed in the majority of the markets. It’s still okay at the very high end, for CPUs and servers, but in all markets, managing cost is really critical.”

“On top of all of that, we have to continue to deliver on time. Because of the complexity, things aren’t getting slower. We’re doing everything we can do continue to keep the same pace as we used to,” he added.

Kengeri said continued advances mean changing the way we think about innovation. It will require continued technical Innovation (materials and processes, device architecture and design-technology co-optimization), but – perhaps more importantly – it will business model innovation. This includes new thinking about long-term R&D focus/ investment, shared investments/learning
euse, and consolidation and collaboration.