Master and Server
Andy Bechtolsheim helped invent network computing. Now, back at Sun, he wants to reinvent it.
By G. Pascal Zachary

(Business 2.0) – Who says you can't go home again? Andy Bechtolsheim is trying to prove that he can--and teach new tricks to the company he co-founded, where he first made his mark as a computer designer.

While a student at Stanford University in the early 1980s, Bechtolsheim designed what would become Sun Microsystems's first workstations, which were popular among engineers and scientists for their speed and powerful networking capabilities. Bechtolsheim left Sun in 1995 to launch a telecom equipment company, which he sold to Cisco Systems in 1996 for stock worth $220 million. He later formed a startup, Kealia, where he was developing high-performance servers, using Advanced Micro Devices's Intel-compatible Opteron chip, to deliver video over the Internet. Sun purchased the company a year ago and, in a coup, brought Bechtolsheim back into the fold as chief architect.

Making Sun shine again is a challenge. The company has seen high-level departures, stagnant sales, and a weak stock price since the bubble burst. But Bechtolsheim's credibility is unmatched among the wizards of Silicon Valley. At age 49, he remains the rare designer who masters all aspects of a system, an unusual feat at a time of extreme specialization. And Bechtolsheim's technical insights have helped him strike it rich three times: first at Sun, then at Cisco, and most recently at Google (as the company's first outside investor, he received shares now worth more than $500 million).

That's why Bechtolsheim's latest pursuit--launching a new line of Opteron servers at Sun--is viewed as one of the most exciting developments in the industry in years. The new computers, code-named Galaxy, are expected to set new standards in the server market, burnishing Bechtolsheim's reputation for engineering excellence and, he hopes, lifting Sun from the doldrums. While tight-lipped about his new machines, which are expected later this year, Bechtolsheim met with Business 2.0 to discuss why he returned to Sun, the new rules for computer architecture, the future of the Valley, and what keeps him running so hard.

As an inventor, how do you measure success?

There are many ways: technological, scientific, financial. My own definition is the impact that an innovation has had on customers.

So despite your credentials as one of the premier geeks of Silicon Valley, you define yourself by the market?

I'm really the entrepreneur who identifies the opportunities as well as the best technical approach to succeeding in the market. My real skill is identifying how to make a successful product.

What's the key to a smash-hit new computer?

The simpler the design, the better. This is difficult for some to appreciate because in modern chip architecture, transistors are almost free. People who design chips don't get penalized for using extra transistors. Yet there is a huge hidden cost for this lack of discipline. More complex architectures consume more power. They run hotter. They have more bugs. So the simplest solution that solves a problem in the most elegant way is always something I strive for.

Why don't simpler designs win out?

Usually, simpler does win, because the ability to evolve things is greater. Of course, in PCs, it didn't end up this way. That's because a large, influential company got behind a more complex architecture. That's what happened with Intel. But the power of simplicity is catching up with Intel. AMD has a simpler chip design.

That's the chip powering your new line of servers. Dell and Hewlett-Packard have been seeing huge sales gains in Intel-compatible servers--the kind you're now designing, albeit with AMD chips. After years of pushing its own architecture, why is Sun shifting tactics now?

Sun has been and continues to be very successful with its own microprocessors. But the market for them isn't growing. When Sun chose its own chip, it did not participate in the so-called industry-standard server market, which is where the majority of the growth has been in recent years.

What must Sun do to tap that growth?

A lot of people think these industry-standard computers are commodities and that you can only compete on price and distribution. I disagree. I can see a lot of innovation that other people don't see. That's what gets me excited. That's why I'm here.

Sun has had its ups and downs. Is it going to be around long enough to see your vision through?

Sun will survive--and thrive. Customers who buy servers prefer a level of support that only a few companies can give. Before Dell, the only company to crack the market and remain a force was Sun. Customers want a stable vendor base, which is another plus for Sun. And the market overall is growing again. Customers are upgrading older systems. So the timing is right for a resurgent Sun. But we need to deliver the latest and greatest technologies. If we do, we'll grow.

High-tech companies still seem to be shaking off the effects of the dotcom bust a few years back. Are they poised for a new round of growth?

In the late '90s, there was a period of overinvestment--too much money went into the same ideas. While painful, what's happened since was in many ways healthy. If you look back, in Silicon Valley the best investments are made when people are more cautious. Because, let's face it, just having money doesn't solve a problem.

After all you've achieved as an engineer and an entrepreneur, why are you still working so hard?

I can't imagine doing anything else. One of my problems in life is that I see opportunities in the market and I can't help but try to pursue them. I get driven by a better solution, and then I go after it.

Have you ever wanted to switch gears and design a consumer device?

I have to admit I have a dozen iPods. I put my entire CD collection on them. It's such a better way to store music. I see a lot of opportunities in consumer electronics, but I don't have a channel to bring consumer products to market. Apple has a great channel, as do the Japanese. But I recently went to a Sony store and I didn't see one thing that excited me. In the Apple store, I see lots of exciting things. There's an important lesson here. People thought consumer electronics was a mature space, and many of them missed an important transition. That could happen in computing.

Is Silicon Valley losing its edge?

Definitely not. In computing, the Valley has the strongest human capital in the world. Despite the rise of China and India, despite the rebound of Europe through mobile phones, the right people are most likely still here. And while there is less money than five years ago, Sand Hill Road is more than filled with enough capital for worthy new ventures. And most of the best ideas still come out of the Valley. Google is a good example.

That said, outsourcing is taking a bite out of the Valley.

The experience in India is mixed. Yes, there are a lot of smart people there. They can support existing products well. But there have been fairly few brand-new products coming out of India. I don't see that changing. They are not on the bleeding edge.

So are there new fortunes to be made?

Well, I wouldn't recommend that anyone start a new microprocessor or memory company today. You're too late. But there's a never-ending cycle of what one can do with ever-improving technology. What's exciting, what makes me optimistic, is that we can attack problems today that a few years ago were unsolvable. Out of these new solutions, new businesses, new industries, and new fortunes will surely arise.