The Revival of Silicon Creativity
Why semiconductor companies need to move away from the smaller, faster, cheaper mantra.
by Keith Cowing, 03/06/08
In the 1960's Gordon Moore, co-founder of Intel, made the lofty statement that semiconductor companies would continue to double the number of transistors they could fit on a chip every 2 years. Now known as Moore's Law, this phenomenon has held eerily true for decades. It was also the main driver behind the growth in the semiconductor industry. Semiconductor behemoths like AMD and Intel have been in a relentless race to continue making their products smaller, faster and cheaper. Year after year. In this brutal industry, one slip can mean missing out on an entire product cycle. However, the ability to make profits by only making things smaller, faster and cheaper is starting to falter. Even Intel, the biggest of the bunch, is facing crushingly tight profit margins and has been through several rounds of layoffs in the last couple years. Low-cost manufacturing overseas has made it all but impossible to actually produce silicon in Silicon Valley, and even cheap labor is not enough to bring profits to an increasingly commoditized business. Semiconductors have been dubbed the crude oil of the technology world, because nothing can run without that processing fuel at the core. So there will always be a demand for processors and communication chips. But if semiconductor companies want to bring back their prominence in the tech arena, it’s going to take more than smaller chips that do the same thing faster.
Physical Limitations of Today's Technology
Computer chips are made by manufacturing silicon wafers with layers and layers of transistors. These wafers, complicated networks of wires and logic gates, produce the computing power that drives electronic devices. The key is the use of a high-precision laser to etch transistors into the silicon. The looming problem, which people have seen coming for awhile, is that the thickness of these transistors is starting to approach the wavelength of the light used to draw them. The inherent nature of lasers makes it physically impossible to carve out transistors significantly smaller than what people are already attempting today. This has led to the strong push in biocomputing and nanotechnology to break through this barrier. However, as of today neither of those fields are mature enough to start feeding the semiconductor food chain. Semiconductor companies are constantly attacking their designs and optimizing their chips, but the law of diminishing returns is starting to kick in.
“No More Moore”
The giant companies that focus on processors and memory (AMD, Intel, Micron, Samsung, etc.) have no choice but to constantly improve chip performance and size. But there is another movement, more relevant to the medium-sized semiconductor suppliers, that is declaring war on Moore’s law and focusing on products that rely on versatility, creativity, and ease of use. Cypress Semiconductor, for example, is promoting an entire “No More Moore” strategy. Cypress is refusing to continue investing in ridiculously expensive manufacturing facilities, leaving this to outsourced partners like TSMC, and is focusing on creating inexpensive products that work in a variety of consumer-oriented devices. An Apple Video iPod teardown highlighted a Cypress chip controlling the click wheel. Other applications range from cell phone touchpads to electronic toothbrushes to vehicle sensors. The idea is to give engineers everything they need to quickly create a new product. Instead of feeding giant telecom companies or PC companies, Cypress is now attacking the long tail of the electronic design world and proving that it can be profitable. In this realm, functionality and versatility are greatly emphasized over pure processing power or transistor size.
Going Mobile
Servers, personal computers, and telecommunications equipment have historically driven a lot of the demand for semiconductors. These areas aren’t going away, but these markets are fairly mature and the room for growth is limited. Mobile, however, is still a wide open field. Cell phones are adding functionality by the minute and evolving into computing machines of their own. The question now is, How can a semiconductor company create a product that solves all the media needs of a phone without draining the battery every few hours? The drivers are still software and functionality-driven, Google’s Android platform, text messaging, blogs, the mobile web, etc. But behind it all are some very versatile chips that need to constantly add functionality. With the proliferation of smartphones (iPhones and Blackberries have become basic life necessities for many people), along with China and India adding cell phone subscribers by the millions, mobile growth is going to be here for awhile. We’ll see who wins the war, and my money is on a player that does more than offer a chip that is smaller, faster, and cheaper.
Image courtesy of FreeDigitalPhotos.net