Online Extra: Intel's Twist on Venture Capital

Its VC arm is less concerned with investment returns than it is with bolstering companies that have hot, chip-hungry technologies

On most planets, losing $1.1 billion in three years on your company's investment portfolio would be a ticket out the door. Not at Intel Capital, the venture arm of Intel Corp. (INTC ). Investment markdowns have taken a bite out of earnings for three years in a row, but most of the same execs remain in place. "A lot of companies went into venture capital as a fashion statement," says Intel Capital President John H.F. Miner. "We have a fundamental belief in our ability to bring strategic value to the company."

Translation: Intel Capital's real job is to help sell chips, not rack up investment returns. The idea is that if certain technologies take off because young companies invent cool new gear, a lot of that gear will need Intel semiconductors. Even if the gear doesn't, it'll boost demand for more powerful PCs and servers that run on Intel chips.

Of course, gauging the extent to which Intel Capital drives technology change is more art than science. Still, Intel Capital seems to have done a pretty good job in pushing forward new technologies that ultimately benefit its parent. "They have a place, especially when it comes to creating standards," says Geoff Yang, a partner at Redpoint Ventures. "Having someone at a high level inside Intel promoting a company is a real positive."


  The best historic example is Intel's late-1990s push to support the then-nascent Linux open-source operating system. Though other tech titans, notably IBM (IBM ), also gave Linux powerful backing, Intel's bets began with 1998 investments in companies such as Red Hat (RHAT ) and VA Software (LNUX ). They gave Intel a boost that lasted long after Intel Capital cashed in its gains from the pair of moonshot IPOs.

Linux now accounts for about 15% of the server market, up from 2% in 1998, and Intel products, led by its Xeon chip, have 86% of the chip market for Linux servers, according to figures from tech consultancy IDC. More recently, Intel Capital's push to invest in Wi-Fi wireless-data companies helped drive fourth-quarter sales of Centrino chips to $400 million, even though the product came out only last March, estimates Morgan Stanley analyst Mark Edelstone. "They absolutely helped," he says.

What's next? Intel Capital is looking for young companies working on technologies for even more powerful wireless communications and the digital home. Its new $200 million Digital Home Fund has put money into Musicmatch, a software company whose products organize digital music collections, and BridgeCo, a Swiss chip company whose products link PCs with TVs.


  It hasn't yet placed any bets to support Intel's push into WiMax, a wireless technology that delivers higher data speeds over a longer range than Wi-Fi. But don't count on it waiting long: Intel plans to begin rolling out WiMax chips later this year, first for communications gear and later for PCs and other devices. "We need external innovation to make our innovation valuable to customers," Miner says. "Our dependence on external innovation didn't go down with the economy."

Neither did Intel's willingness to open the checkbook to support a new idea. The 200-staffer venture arm invested $700 million in 120 deals in 2003. Almost $575 million went into Micron Technology (MU ) and Elpida Memory, two companies developing next-generation memory chips. Intel needs faster PC memory to clear a potential logjam in delivering movies and other digital content in PCs and home networks. If consumers like the idea, Intel will cash in as more homes add second PCs.

In good times, Intel Capital does make money. Over its 13-year history, it has earned $3 billion, about enough to pay for a new chip plant. And that may be just the beginning. If Miner's latest plans work, Intel Capital will help sell lots of new chips -- and cash in a few, too.

By Timothy J. Mullaney in New York

Before it's here, it's on the Bloomberg Terminal.