Michael Fister left a 17-year tenure at Intel to take the helm at San Jose, Calif.-based Cadence Design Systems, an electronic design automation (EDA) company that produces software and hardware methodologies used to design and verify advanced semiconductors. Fister recently talked with CE about evolution in the EDA industry and computer power.
How is the approach to EDA shifting?
There’s a perfect storm of things coming together-impossibly complex physics, complex devices, short time-to-market cycles, automation-right now. When I was a young man, five of us would work on the chip. Now hundreds of people work on the most complicated chips in the world.
There’s no one person who understands the totality of what’s going on-just as someone who manages a $1 billion corporation doesn’t know every detail about everything.
You left Intel to join Cadence-what do you view as the company’s competitive strengths?
We bring a breadth of technology, as well as geographic coverage. Early on while I was at Intel, you had to buy one capability from one company, a different one from another company, and yet another from a third company because they were all very point-tool-oriented. First, that makes it hard to integrate and optimize between those pieces, and second, none of the players have a view of the total problem you’re trying to solve because they’re all too point-based. So one of the advantages we offer is an end-to-end dynamic, what we refer to as a consolidating phenomena. When you can show that you have that technology breadth, geographic coverage and financial security, you become a dependable partner. That’s an earned respect that involves a commitment both ways.
How do you ensure you bring enough value to justify the fees you charge?
Computer design must be rooted in the practical. So much of the history has been, “Here’s an impossibly complex software tool that you couldn’t possibly understand, but please pay me for it, and I hope it’s good for what you do.” I used to call that YoYo, “You’re on your own. Use my tool, pay me, and let me see what it is.”
When you develop partnerships, you learn together, evolve and depend on each other. That’s what made the IT industry move and made Microsoft the stalwart of operating systems or Oracle of database technology. Oracle used to say things like, “Mine’s better, because mine’s object-oriented or mine’s relational.” Now they say things like, “I want to help you solve customer relationship management for your business,” or “I want to help you with a people application.” That’s what we’re trying to do in our industry, which is still a little bit stuck in the “my technology’s better than yours” mode.
Are Chinese or Taiwanese firms trying to give you a run for your money?
You can count on one hand the number of Chinese EDA companies.
How long will that last?
I don’t know. Maybe not very long, maybe for a long, long time. We’re trying to ensure the second with a solutions-oriented approach, because that’s a natural value for us as a broad line supplier. It’s very difficult for someone to start that type of business. I’m not saying that there won’t be some kind of local capability develop in China. If there is, we’ll compete with it. Most of what we see is the opposite of that.
When I visited China at the end of last year, there were over 800 fabless semiconductor startups (companies that contract out chip production). There are a lot of people out there starting businesses to try to figure out how to build a special integration or a simpler add-on device. We’re an enabler to that, because we give them the tools, reference examples of how to do it, and even help them use the toolbox. In a sense, we’re unlocking Chinese evolution.
Where is computing power going?It continues to grow and grow. We are essentially a facilitator of computing power, because anybody who builds a semiconductor probably uses some or a lot of Cadence software. It’s just kind of the way the world works when you are a market leader in this racket.