Company Builder

Take the Hard Road

RFI is situated in an unfoundrylike corporate office building on North First Street in San Jose. There's nary a tattoo or a pierced body part in sight. Rather, RFI is staffed by grown-ups -- an ethnically diverse group of veteran engineers and businesspeople who have led startups and worked in senior positions at such global heavyweights as Cisco, IBM, Intel, and National Semiconductor Corp.

At 8 AM, Atiq Raza dashes over to his scribbled-on whiteboard and quickly erases names and numbers for his next round of funding. Outfitted in a blue Oxford shirt and a pair of khakis, Raza instantly evinces a technophile's love of gadgetry. He is a power user of PowerPoint. He flashes a laser pointer. He fires up video clips from his Sony Vaio notebook computer. This is a performance in the guise of an interview: Raza speaks in complete paragraphs, and he's determined to make his case.

The first question hits a sore spot: What's the difference between RFI and all of the Internet incubators out there today? Raza avoids the "I" word as much as possible. Incubators have been hammered since dotcom fever broke this past March. Besides, Raza displays an engineer's contempt for the incubator model: It's all too easy, and there's no barrier to entry.

"Companies in the dotcom world are too simple to form," he says. Warming to his theme, Raza grabs a sheet of paper from his desk and recites a rundown of some of the portfolio companies at Internet incubator CMGI: CarParts.com, Dialpad.com, GoFish. com, MyFamily.com, Snapfish.com.

"Not our cup of tea," he sniffs.Why not?"Because these are epiphanies. They might be great ideas, but the ideas define the company. In our case, a process defines the company -- a process that captures the combined knowledge of the founders of the company, the experts at RFI, and a network of outside experts."

Customers Define Your Product

The first stage in building a company, says Raza, is to crystallize the company's product definition. It doesn't have to be cool. But it must be compelling. "A compelling product is something that target customers really, really want," explains Raza. "It's something that solves an economic problem." And how do you ensure that your potential customers need your Next Big Thing? You get them to help define your product.

Such was the case with YuniNetworks. By all accounts, the Yuns' pitch meeting with Raza went extremely well. "Ken and Kay were phenomenal," Raza recalls. Still, he was troubled. The prototype for Yuni Switching Fabric, developed by Ken Yun and his team of graduate students at UC San Diego's computer labs, hadn't quite jelled. So Raza and the Yuns took the prototype to a potential customer (a company that Raza declined to name).

They flew to a remote site, where the un-named company's R&D labs are located. Raza and Ken Yun outlined the technology for the company's network architects, who proceeded to "rip the plan to shreds."

"But the top manager told us not to get discouraged," says Raza. "He said that this was a critical technology, and since it was important to his company, his team would help us integrate it with the rest of the system. I told him if that's the case, then we'd deliver this product faster than he'd ever seen a product delivered."

Raza returned to San Jose and called a leading network architect at another company that was a potential customer. In one four-hour session with Raza's team of engineers, the second architect made great progress in mapping out a solution to the technical problems that had already been identified. Many more meetings took place over the next several weeks until the "product definition" had crystallized -- that is, until the group had forged a blueprint for a switching fabric with a clean interface. And that was Raza's great innovation: By getting the two companies to help YuniNetworks spec out a faster, better product, Raza locked in two of YuniNetworks's biggest potential customers -- before the company's product hit the market.

Use It and Reuse It

According to Raza's worldview (and he certainly is not alone in this), communications has become the new frontier for the semiconductor industry. The increased power of desktops and servers has put huge demands on the world's computing infrastructure. "The economic value has moved from the desktop to the infrastructure," says Raza. "There has been a huge migration from the microprocessor world into the broadband-networking world." RFI is focusing on that world with laserlike intensity. If it isn't broadband, Raza isn't interested.