The Constant Gardener
At Kleiner Perkins, he nurtured seedlings that grew into the Valley's mightiest oaks. Now, out on his own, Vinod Khosla is tilling the startup fields again.
By John Heilemann

(Business 2.0) – Vinod Khosla is often described in terms of the scale of his ambitions, and his abject refusal to acknowledge impediments that others consider insurmountable. The first time I ever heard him speak--nearly a decade ago at a private retreat in Aspen hosted by his firm, venture capital powerhouse Kleiner Perkins Caufield & Byers--he was urging his partners to finance an alternative global telephone system. I later described him in print as "wild-eyed," although "lunatic" would have been equally apt.

So I wasn't surprised by Khosla's reply when I asked him recently, at the urging of my editors, for his essential golden rule of business: "Since one fails often, address markets that make it worthwhile when one does succeed." (For others' golden rules, see page 108.)

In his 20-year career in venture capital, Khosla has indeed failed on multiple occasions, sometimes spectacularly. (Dynabook Technologies, Go Corp., and 3DO leap instantly to mind.) But his successes have been so large and influential that they've made his missteps seem trivial. A founder of Sun Microsystems, Khosla emerged in the late 1990s as one of Silicon Valley's most prodigious VCs. With a trio of investments in networking startups--Cerent, Juniper Networks, and Siara Systems--he produced returns of more than $4 billion. And that was on top of Excite, NexGen, and a panoply of other hits. Together with his friend and longtime co-conspirator John Doerr, Khosla transformed Kleiner Perkins from the king of Sand Hill Road into something more mythic: the venture firm that financed the greatest boom (and bubble) the world has ever seen.

Earlier this year, however, Khosla stepped away from Kleiner Perkins, declining to sign on as general partner in its latest fund and striking out on his own. And while he's headed now in a surprising new direction--about which more in a moment--Khosla's stance remains the same: If the market's not vast, the problem's not hard, and the solution's not elusive, then he's not interested.

Around the Valley, there's some confusion about what exactly Khosla is up to. When we met in October at the Web 2.0 conference in San Francisco, scuttlebutt was whipping around the hall that he was raising a new fund--a potential Kleiner Perkins rival. But Khosla quickly quashed those rumors. "I am investing my own money on my own account in an informal way," he said. "I'm not starting a fund."

He's hardly pulling back from the fray, though. At 50, with four kids, Khosla remains an acute observer of all things new and spiffy. Before we got together, he sent me an epic PowerPoint deck titled "Re-Inventing Forms," replete with analysis of the innovations that he sees propelling the Web forward--from BitTorrent, VOIP, and Wi-Max to collaborative filtering, mobile content, and "open-source everything," including textbooks.

A few years back, during the bust, Khosla toted around with him a different deck, one with a section named "The Illiquid Bulge." The phrase referred to the thousands of boom-era startups that had yet to go public, get acquired, or go under--and to the heaps of VC money raised during the bubble that had yet to be invested. At the time, Khosla believed that, until the bulge was digested, the Valley would remain on pause. "It might take until the end of this decade before we see another boom," he told me in 2003.

With signs of a new boom everywhere, Khosla says now that the first part of the bulge has dissipated. "Some companies have gone out of business, but others--the creative ones--have found new ways to survive," he says. But Khosla worries that there's still too much venture capital floating around the Valley. Thus he has a "word of caution" for this generation's entrepreneurs: "Just because you get funded does not mean you are successful."

Khosla himself has always been a famously tightfisted backer. More than half of his investments during the boom involved sums of less than $1 million. Khosla explains, "My friend Mark Leslie, founder of Veritas Software, says that the more money you give a company to start with, the less likely it is to be successful. The more money the founders have, the more confident they get about their business plan, the less they experiment." He goes on, "The right way to build a company is to experiment in lots of small ways, so that you have plenty of room to make mistakes and change strategies."

Since splitting off from Kleiner Perkins, Khosla has dropped his cash into a handful of firms that hew closely to those principles--most developing patentable technology, not consumer Internet services. There's Beceem Communications, a Wi-Max chip outfit. There's Xsigo Systems, a data-center equipment maker, in which he's invested alongside Kleiner Perkins. And there's Spatial Photonics, an outfit designing new display "engines" (ditto).

But while Khosla clearly believes in these companies, it doesn't take long to realize that they're not where his passion lies. Instead he's obsessed by a crusade as grandiose as any he's ever pursued: ending America's dependence on foreign oil. "I think we have a replacement for oil today," he says. "It's cheaper, cleaner, it doesn't require a change of infrastructure, and it appeals to most of the lobbies. What is this platform? It's ethanol."

Ethanol has been around forever, but Khosla points out that new technology is transforming its utility. "In the past, ethanol was made from corn, which isn't that great environmentally and isn't very efficient--for every one unit of energy you get 1.5 units of fuel," he says. "Now, with bioengineering, we can make ethanol from agricultural waste--which is four to eight times as efficient." He believes that, with the right incentives, automakers could be induced to create "flex-fuel" vehicles that run on either ethanol or gas or some blend of the two. (There are already 4.5 million such cars on the road in the United States today.) Both the farm and environmental lobbies would favor the switch, and so would consumers, because ethanol is cheaper than gas. Within five years, Khosla argues, the transformation could be largely complete.

Khosla isn't the only VC in the Valley with his sights set on alternative energy. The area has been trendy for some time now, with fuel-cell startups, in particular, receiving considerable investment interest. But Khosla's ideas, more than most, seem to have caught fire with policymakers in Washington. After he gave a talk about ethanol at the Clinton Global Initiative conference in New York in September, Khosla was mobbed by politicians including Al Gore and former Democratic Senate leader Tom Daschle. And on a recent trip to the capital, he received a warm reception from quarters ranging from the tree-hugging left to defense-minded neoconservatives.

On the surface, Khosla's interest in ethanol might seem a bizarre tangent. Yet it fits snugly with his long-held tendencies: Here you have an enormous problem, but one where miracles of science may make the intractable tractable. And here you have a putative revolution that could lead to a financial jackpot. As Khosla sheepishly admitted to me, "I started on this from the environmental angle, but hey, it's a great investing area. As soon as you're talking about biotech, you can start a company, no problem."

If Khosla's ethanol solution sounds too good (and too easy) to be true--well, it may just be. But the same was said about Sun's workstations in the early 1980s and about much of the optical networking gear Khosla financed in the 1990s. At Kleiner Perkins, he constantly fought an uphill battle to convince his more sober-minded partners that his flights of fancy were more plausible than they seemed. The question was whether those partners served a valuable gating function, preventing him from diving off any number of cliffs, or were shortsighted naysayers who kept him from achieving even more than he did. If, five years from now, we're all driving cars that run on switchgrass, we'll finally know the answer.