Failure Is an Option

An impeccably dressed man careened through San Francisco's Union Square, muttering incoherently, weeping without concern for the stares of tourists and shoppers. Just moments earlier, before an Olympian gathering of financial analysts and investment bankers, he had made what he had thought was the speech of his life. In reality, it had been a final, desperate plea for the life of the Silicon Valley startup that he ran. But his words had yielded mainly indifference -- certainly nothing in the way of life support. And that reaction to his words had yielded, in him, a moment of shattering clarity.

Rick Shriner found his car and headed south on U.S. Highway 101. "It's over,'' he repeated to himself, choking on his tears. In his head, he rehearsed what he would say to Gordon Campbell, a legendary Silicon Valley entrepreneur and the chairman of his company's board. Shriner also began composing his next speech, the one in which he would tell his team at Exponential Technology Inc. -- more than 70 engineers, managers, and staffers, who had burned through nearly $30 million in capital and through four years of effort, much of it measured in 15-hour days -- that their dream of building the world's fastest microprocessor was over.

As he drove south, Shriner cursed Apple Computer -- his startup's first outside investor and its most important potential customer. Apple had given Exponential life; now, in the wake of a dramatic repositioning, Apple was helping to end that life. And he cursed himself, engaging in what-might-have-been speculations. What might he have done to postpone this day of reckoning? How had his miscalculations contributed to Exponential's demise?

"I let myself believe that we were onto something really great," Shriner, 55, says today. "I told myself, 'If we can just keep going, we'll make it.' That was a big mistake.''

Reality bytes.

Not every Silicon Valley startup becomes Cisco or Yahoo! or Sun Microsystems. Short lives, followed by quick flameouts, are the more common trajectory. Plenty of startups fail because of ineptitude. The technology doesn't work; management doesn't execute. That's not true in this case. Exponential was a small (and perhaps unavoidable) casualty of one of the biggest comeback stories in recent business history: the rebirth of Apple Computer. Today Apple is enjoying fast growth, healthy profits, and a surging stock price. But all that's left of Exponential is a forlorn sign on an office building in San Jose -- and, naturally, a massive bout of litigation. (Exponential has sued Apple for $500 million; Apple has countersued for breach of contract.)

What's so instructive about the Exponential saga is that the company did many things right -- but failed nonetheless. It failed mainly because the odds of success for any startup are incredibly long. What's so poignant about the Exponential saga is that many of the characters in it are grateful for the experience -- and would even go through it again. "As a parent," says Paul Nixon, 41, a microprocessor engineer who left Apple to join Shriner's outfit, "you teach your kids to play hard and, if they scrape their knee, to rub it, clean it up, and play the game again. My kids got to see me doing the same thing. That's how you learn and make progress. Our fight for survival was like a PhD program in team dynamics, leadership, and high-tech business strategies."

Startup

"How would you like to have the world's fastest chip?"

Startups that aspire to greatness begin by identifying a big market opportunity. Exponential's opportunity was big -- but it was also rooted in a power struggle among several big companies. In 1992, Apple, IBM, and Motorola formed an alliance (called AIM) to make one last-ditch effort to turn back the Microsoft-Intel juggernaut. The goal of AIM was to produce a microprocessor that would set a new performance standard for the industry. The name of this device was the PowerPC, and its mission was to drive a series of personal computers that would be so blazingly fast that consumers would give the Macintosh operating system (Mac OS) one more chance against Windows.

Unfortunately for AIM, Intel did not stand still. Rather, Intel chips just kept getting faster. Gordon Campbell, 54, founder of two well-known semiconductor companies, was aware of Apple's plight. He and two engineers -- George Taylor, 43, and Jim Blomgren, 38 -- began kicking around new ideas for hyperfast chips that would run the Mac OS. They proposed a design that would use BiCMOS technology, which was ultrafast but which burned too much power and expended too much heat. In the spring of 1993, while Taylor and Blomgren addressed engineering challenges posed by BiCMOS, Campbell assigned a finance specialist named Stephanie Dorris, 36, the task of drawing up a business plan for a new company, which he called Renaissance Microsystems.