What follows is an excerpt from a speech I gave last week at an event in Palo Alto, Calif.
There's a great metaphor for the change that's happening today: You're standing in it. This Four Seasons Hotel and the building next to it are but a few years old. The small area that these buildings inhabit was a rundown slit of land called Whiskey Gulch. If you were traveling southbound on 101 and wanted to get into Palo Alto, you slid off the highway into this place, which had the look and feel of an old Wild West town. There were pawnshops, liquor stores, bars and secondhand-clothing shops, and any number of people out front who had fallen on hard times and couldn't get up.
In a matter of a few years, that social intractability was bulldozed away and replaced with a luxury hotel and a building housing a leading intellectual-property law firm.
Change comes on hard when it comes. Your industry has been forever changed, and the only thing that's left is for the dozers to roll over the companies that still live in Whiskey Gulch. This isn't about new cycles or smoother cycles in a traditionally cyclical business. It's about a different business.
Whether it was the mainstreaming of the PC or the mobile, expectations were set among average consumers in the past two decades: Every year or two, something better, cheaper, cooler would come along. Did I say it needed to be cheaper? Sure. Semiconductor average selling prices peaked around 1994. They've slid ever since and are never coming back.
The second thing that happened was just as fundamental. The communications infrastructure was built out and really started to enable the Internet. Then Web 1.0 blew up in a lot of people's faces. There was much gnashing of teeth. The End was Here.
Well, that was like predicting Willie Mays would be a failure because he went 0-for-21 in his first major-league at-bats. As if in step with the history of California--where if something falls down or burns down or gets washed away, you start rebuilding the instant the rubble is cleared--Web 2.0 was born. It has torn down barriers to entry and is changing the face of business.
Technology the way we've defined it has ceased to be as important as it once was. Innovation has shifted to business practices and software--both embedded and consumer applications. That's where the real innovation is happening.
It's estimated that to get a return on investment at the 65-nanometer node, an IDM must generate $8.3 billion in sales per year. At 45 and 32 nm, it must generate $13.3 billion and $16.7 billion in annual sales. So there you have it: The IDM model is dead.
The business is shifting in other ways. Given the maturity of the industry, do you really need six FPGA companies? Ten microprocessor core vendors? How many analog companies do you really need?
The notion of building a company now is radically different. It can cost anywhere from $50 million to $75 million to design a system-on-chip. Never mind that you need a big end market to tap into to make those numbers work: Sand Hill Road isn't dropping $50 million to silicon startups.
The revolution that engineers brought to the world during the past 50 years is now showing its full weight, and it's head-spinning. You have enabled a technology ethos and infrastructure that are revolutionizing business and the definition of value. And it's all starting to turn around and affect the industry that wrought it.