(FORTUNE Magazine) – The moment was vintage Craig Venter: Biology's bad boy stood before a crowd of reporters in Washington, D.C., trumpeting his latest achievement, with a beaming Spencer Abraham, the U.S. Secretary of Energy, by his side. Venter announced he had assembled a virus--a harmless, bacteria-infecting virus--from scratch, in his lab, in two weeks flat. The only other lab to attempt such a feat had taken three years. What was more, he went on, this was no mere test-tube sport. It was a small but important step in a much larger quest.

Venter, in case you don't remember, is the renegade scientist who in 1998 challenged the U.S. government in a race to sequence the human genome--and fought his giant rival to a history-making tie. He's the same guy who was later ousted by the genomics company he had founded after failing to deliver on outsized promises.

Now he has shifted to what may be the ultimate scientific ambition: To create new forms of life. Piecing together the virus was just a dry run. Venter plans to use commercially available snippets of DNA to assemble the smallest and simplest of genomes, a lab-made thread of just 300 genes, and insert it into a bacterium that has been stripped of its own genetic code. Then he'll hold his breath and see whether the tiny microbe "boots up" and starts moving, metabolizing, and multiplying, all according to Venter's reductionist version of life's playbook. If the plan succeeds, a new creature will have entered the world. And Venter will have taken a bold, juicy bite of what some see as the forbidden fruit.

He has big plans for his designer microbe. He calls it his model A and intends to use it as a chassis on which to build a fleet of bigger, more complex bacteria crafted to pump out pharmaceuticals and chemicals more cheaply and efficiently than they can be made today. Other bugs built on Venter's minimalist chassis would help reduce humanity's environmental and energy woes. They would capture carbon dioxide, derailing the greenhouse effect. They would clean up nuclear waste. They would generate hydrogen in quantities that could reshape the global energy economy. And as Energy Secretary Abraham attested at the November press conference, Uncle Sam is on board --the Department of Energy has backed Venter's venture with $12 million of taxpayers' money. "These dreams are not science fiction," Abraham said. "The potential for this research to revolutionize our future is enormous."

Venter's critics quickly came out in force. The press conference was a "sales sound bite," charged Eckard Wimmer, a virologist at the State University of New York at Stony Brook, whose three-year virus-building effort Venter had portrayed as the slow boat to China. "Science by press release," snorted Richard Ebright, a Howard Hughes Medical Institute investigator at Rutgers University, adding that Venter's virus synthesis "could not by any conceivable stretch of the imagination" be described as an important advance over current commercial technology. George Church, a Harvard Medical School geneticist who, like Venter, is working on synthetic genomes, added with amusement, "Craig does have an excellent PR machine."

Thus begins act two for J. Craig Venter, gene sequencer extraordinaire. Nearly four years have passed since his triumph at a White House ceremony in which President Clinton acclaimed the deciphered genome as the "most important, most wondrous map ever produced by humankind." Nearly two years have passed since Venter was ignominiously dumped as CEO of Celera Genomics, the gene-sequencing company he founded. Now, at 57, the guy who is arguably the world's most ambitious scientist (which is really saying something) is revved up again. "This [project] potentially will replace the entire pharmaceutical and chemical industry," he says.

The truth, as is almost always the case with Venter, lies somewhere between the hype of the press conference and the harsh charges of detractors. Venter has indeed found a clever, speedier method of melding small pieces of DNA into long strands. With refining, it will certainly help toward the eventual goal of stitching together a genome for his "model A" microbe. But did that warrant a press conference? Or justify the Energy Secretary's toasting Venter's work as a "critical accomplishment"?

Before we go further, let's pause to clarify: Venter isn't literally trying to create life in a test tube, not the way you've seen in science fiction movies--zapping a bunch of inert chemicals to produce, say, a silicon-based life form that breathes ammonia, basks in X-rays, and eats rocks. His minimal-genome quest is more akin to cloning--the cobbled-together DNA he inserts in a bacterium will depend on that bug's own cellular machinery to work. But by pushing the frontiers of a new field called synthetic biology, he's close enough to creating life in the lab to raise both mad-scientist fears and a host of safety concerns--of freakish new bugs that might infect people, or fall into the hands of terrorists, or escape to destroy the environment. We'll come back to that later.

Harnessing bacteria isn't as wacky or futuristic as it may sound. Venter and others are engaged in a kind of fast-forward microbial equivalent of the work our remote human ancestors did in domesticating birds and mammals. And industrial scientists have been tinkering with bugs for decades--the agriculture and biotech industries use bacteria to make proteins, vaccines, and vitamins. But now microbe manipulation is moving into the 21st century. Last year, for example, University of Massachusetts biologist Derek Lovley used a common bacterium called Geobacter to clean up uranium-contaminated groundwater. The DOE paid for that work too.

A bug like the one Venter wants to create, with the slimmest of genetic instruction books, would be immensely valuable commercially. Today's industrial-duty bacteria come freighted with all kinds of genes not relevant to human purposes--such as backup systems evolved over eons to help the microbes survive in the wild. That genetic baggage consumes energy; remove it somehow, the theory goes, and you can make industrial microbes far more efficient.

The mainstream approach to streamlining genomes is more straightforward--and less scientifically risky--than what Venter is attempting. It is to whittle down the genes of existing lab and industrial bacteria. University of Wisconsin geneticist Fred Blattner has formed a private company, Scarab Genomics, to do just that. He has produced a stripped-down version of the scientific workhorse bacterium E. coli. Blattner hopes that his pruned-down "clean genome" bug will prove its safety and efficiency in the lab and in industry. Unlike Venter, Blattner says, "we are heading directly to the market with something that is technically feasible now." Scientists at MIT are also working on pared-down versions of existing bugs.

But Venter and other synthetic biologists say that approach is shortsighted. It confines scientists to working with the tiny fraction of bugs that will grow in the lab--a little like a drunk on the street at 2 A.M. hunting for his lost car keys under a lamppost, because that's where the light is.

Instead, the synthetic biologists propose to build genomes and create useful bugs from scratch. These new life forms--call them synbugs--will hugely expand scientists' power. By building into the synbug genes collected from the natural world, scientists will be able to replicate the capabilities of millions of "wild" species that can't be grown in the lab. A handful of prominent researchers at places like Harvard and MIT are working on genome synthesis, but Venter is the only one so far who is daring to promise world-changing results.

But that's Craig.

He calls himself a "superenzyme": a guy who, like a chemical catalyst, takes a slow-as-molasses situation and speeds it up by orders of magnitude. That aggressiveness is evident as he parks his black Porsche (in a no-parking zone) in front of his state-of-the-art DNA-sequencing facility in Rockville, Md. Put together in less than a year and outfitted with 100 ultrafast gene-sequencing machines, workspace for a staff of 90, and a heavy-duty data center to crunch all the information that the sequencers generate, it exceeds the technological muscle he commanded at Celera. It's one of perhaps three sequencing facilities of its caliber on the planet. For the project he has brought with him Hamilton Smith, the Nobel laureate who was Celera's chief scientific brain. Apart from the DOE grant, the project is running on Venter's money--to fund his scientific ventures he established a $100-million-plus endowment from his Celera stock profits and those from an earlier partner company, Human Genome Sciences.

Venter has penetrating blue eyes and a bald, sunburned head that attests to a love of sailing and sunny climes. He moves with the energy of an entrepreneur 30 years younger, and comes across with a mix of braggadocio ("This is a $30 million room," he volunteers during a tour of the sequencing lab) and self-deprecation ("At least some people like me," he cracks about the collection of awards in his office).

He was born in Salt Lake City just after World War II to an accountant dad and homemaker mom who had met in the Marines. Soon they moved to Millbrae, Calif., where they began raising their four kids in a bungalow under the roar of planes landing at nearby San Francisco International airport. Craig, the No. 2 child, grew up in the shadow of his older brother, Gary, who is now an insurance risk analyst. At high school in the early 1960s, Venter set records on the swim team but nearly flunked out; he had no plans other than to chase girls and surf. Says his younger brother, Keith, now a NASA architect: "He just wasn't motivated academically. That stuff was just so unimportant to him."

Vietnam changed that. Driven by horrifying experiences he had as a Navy medic at a field hospital in Da Nang, Venter zoomed to a bachelor's degree and a Ph.D. in physiology and pharmacology in six years. Later, as a research scientist at the National Institutes of Health, he pioneered a technical shortcut that mightily hastened the hunt for genes. He and others used the method over the next three years to discover and decode tens of thousands of human genes.

In 1995, Venter became the first to sequence a whole bacterial genome, decoding the DNA of a bug called H. influenzae that causes pneumonia and often lethal childhood meningitis. In doing so, he leapfrogged rivals and flummoxed officials at the National Institutes of Health, who had refused to fund the project on the ground that the novel method he wanted to employ --known as whole-genome shotgun sequencing--wouldn't work. (Whole-genome shotgun sequencing is a brute-force process in which all of a cell's genes are shredded into millions of DNA snippets to be read, or "decoded," by sequencing machines. Computers take the resulting morass of data and piece together the bits of code into sequences of whole genes. Imagine chopping 20 encyclopedias into a mountain of confetti, then reassembling the words, sentences, chapters, and volumes.) James Watson, the Nobel laureate who co-discovered the double helix of DNA and who is no doubt the most distinguished of Venter's many critics, matter-of-factly acknowledged his achievement as a "great moment in science."

Venter soon became intrigued by a big question: What is the minimum set of genes necessary for life? Humans, animals, and plants have tens of thousands of genes, but single-celled creatures like those Venter had begun sequencing have relatively few. In 1995 a team led by Venter's wife, Claire Fraser, herself a distinguished biologist, decoded the smallest-known genome of a free-living organism, a parasite called M. genitalium that infects the human urinary tract. It has just 470 genes. Venter teamed with Fraser and others to explore how far M. genitalium could be trimmed. Just as Venter shifted his attention to the human-genome race, the work led to a surprising result: Of the bug's 470 genes, only an estimated 265 to 300 were essential to life.

But the world did not know of Craig Venter until he took on Uncle Sam. In 1998 he announced the formation of Celera and its intention to sequence the human genome, start to finish, in three years. The firm was funded with $330 million from its parent, lab-machine maker Perkin Elmer, and equipped with 300 of the newest, fastest gene-sequencing machines, made by another Perkin Elmer subsidiary.

At that point scientists in the government's Human Genome Project had spent eight years mostly doing preparatory work and had sequenced just 3% of the genome. They hoped to finish the $3 billion project by 2005--a deadline that seemed in real danger of slipping. Within a month of Venter's jump into the race, though, the Human Genome Project moved up its target to 2001 and federal officials began plowing in tens of millions of dollars more. Two years and buckets of bad blood later, Venter and the government scientists reluctantly shared credit for simultaneously completing their genome drafts. Though both drafts were full of holes--and are still being refined--it was a clear victory of quick-and-dirty science over the slow and painstakingly thorough approach many researchers favor.

Venter's triumph also marked a high point in the annals of egotism. During the genome race there had been rumors he was using his own DNA as the primary specimen for spelling out the human genome. That was completely contrary to what Celera was telling the world. The company made an elaborate effort to recruit a diverse pool of blood and sperm donors. Each volunteer read and signed a 30-page "informed consent" document that had been painstakingly developed by a panel of distinguished scientific and ethical advisors. At the White House celebration Venter touted Celera's multiethnic volunteer pool. "We did the sampling not in an exclusionary way, but out of respect for the diversity that is America," he declared. Having a diversity of samples didn't matter from a scientific perspective, but it was symbolically important for the first reading of the human code.

It was a big charade. Venter later acknowledged that most of the genes Celera sequenced had been his very own. "Craig did what I feared he would do," says Arthur Caplan, a University of Pennsylvania bioethicist who is Venter's friend and who served on the advisory board. "He [used his own DNA] because he felt he had earned it. It was wrong." Caplan ultimately published an indignant op-ed in the New York Times titled "His Genes, Our Genome." Venter says he did it to set an example. Too many people, he says, are leery of undergoing genetic analysis for fear of job loss or insurance discrimination: "The best statement I could possibly make that people should not be afraid to have their code sequenced and seen by others was to do it myself." What's more, he says, analyzing his genetic code indulged a deep, natural curiosity about his body's makeup; he's writing a book about what he learned.

The genome race made Venter not only famous but also rich. Celera went public in the spring of 1999; when the stock ballooned the next year, the value of his 5% stake approached $700 million. And however much his academic rivals hated his hot-dogging, they had to concede he'd made it to the top of their heap too. He ranks in the top 0.05% of molecular biologists, according to the Institute for Scientific Information in Philadelphia, which measures influence by tracking how often scientists' publications are cited by their peers. There was, of course, talk of a Nobel.

Then things began to go wrong. Investors had poured billions into Celera as it climbed from $7.20 a share in mid-1999 to $247 nine months later. But soon they started to realize that Venter had overhyped the company's potential. He had promised to make it the "Bloomberg of biology" with 100 million customers by developing a one-of-a-kind database of genomic information. Celera had considerable success in signing up subscribers--especially big drug companies willing to pay top dollar for an early look at its findings. But there was no way to turn a profit big enough to justify a three-digit stock price. Less than a month after its peak, Celera's stock fell to $92 a share; by the end of 2000 it was at $36. (Full disclosure: This writer was silly enough to buy Celera in early 2000. At a recent $15.41 a share, my $969 investment was worth $122.88.) The end for Venter came in January 2002, when Celera's board unanimously voted him out. He had to creep back into his office after midnight to collect his belongings.

Venter retreated to Saint Barts, where he holed up on Sorcerer II, his new 95-foot yacht, in a deep funk. Shortly before leaving for the Caribbean, he had dinner with James Shreeve, a writer who had spent years at Venter's elbow chronicling the genome race (and whose vivid account, The Genome War, was published in January). "He was not the Craig I had followed around," says Shreeve. "He was devastated. He looked smaller. He seemed unsure of himself in the crowd." Later, on the yacht, Venter told Shreeve, "There is at least as good a chance that I'll commit suicide as die from some disease."

He tried shopping therapy, buying a $5 million villa complete with a beautiful pool with a nine-foot waterfall, high on a mountainside overlooking the Caribbean. But the good life made him super-restless. Says brother Gary, about time they spent together that year at a house Venter owns on Cape Cod: "I'm seeing Craig out on the deck, drinking mojitos and trying to get into being a man of leisure. That would only last an hour ... It seemed to be driving him crazy."

It took just a couple of months for the old Craig to resurface. In April he unveiled the J. Craig Venter Science Foundation. (The $100-million-plus he gave the foundation accounts for more than 90% of the cash he took out of Celera, Venter says.) He also announced he was returning to the minimal genome quest that had fascinated him before the human genome race had intervened. By year-end he had enlisted a small team, led by Nobelist Smith, to tackle the daunting array of challenges involved in creating a synthetic species.

If all goes as Venter envisions it, the little M. genitalium bug he worked on in the mid-1990s will serve as the framework for his "model A." The biggest hurdle is to figure out exactly which 300 or so genes will bring a cell to life. What's more, no one has ever synthesized a genome remotely as large as the one Venter proposes. The biggest virus genome ever built in the lab is about 7,500 genetic code letters long; Venter's model A genome will be 40 times that size, dozens of times longer than the strands of DNA he has learned to produce. Removing and replacing the bacterium's DNA is another major obstacle. And even after that's done, it's unclear whether the bug will "boot up." If it fails to do so, the team may not be able to figure out why, because the processes by which a bacterium executes its genetic code are so poorly understood.

Any of those challenges could derail the project long before Venter can even begin outfitting bugs to produce medicine or capture carbon dioxide from the atmosphere. A failure would confirm the suspicions of many scientists who think that this time Venter has overreached. Take the highly knowledgeable observer who dismisses Venter's work as "tinkering." So-called rational bacterial design--building bugs from the ground up, knowing exactly how they are going to perform when you launch them because you understand the most intimate details of their biochemical workings--is a century away, he says. "You and I won't be here."

Even fans are skeptical. Says Norton Zinder, an eminent molecular biologist who was instrumental in getting Venter elected to the National Academy of Sciences in 2002: "On paper it works, [but] there's more than paper to living material."

Venter is working hard to demonstrate the practical potential of his enterprise. In a recent experiment sequencing the genomes of bacteria in the Sargasso Sea off Bermuda, he found thousands of new species of microbes containing more than a million hitherto unknown genes. That's about ten times the total number of genes discovered in all species to date. He will publish details of his findings in an upcoming paper in the journal Science. Among other things, he says, "we've literally come up with hundreds of new [chemical] pathways and genes that might enable us to generate hydrogen biologically."

Venter is happy to provide other tantalizing suggestions of what may be possible if his synthetic genome succeeds. Consider Taxol, the breakthrough medicine used to treat breast and ovarian cancer. It is derived from the bark of the yew tree; three to six of the large, slow-growing trees are needed to provide the 60 pounds of bark it takes to treat one patient for a year. A semisynthetic form of the drug based on a chemical from the needles of the yew is still costly and hard to produce, and making Taxol from scratch in a test tube is prohibitive. With a minimal-genome microbe in hand, says Venter, "we would start today and go sequence the genome of the yew tree. Get the gene pathways for making Taxol. And put them in the synthetic cell and make Taxol."

He also loves to talk about an exotic organism called Methanococcus jannaschii. It lives at ultra-high temperatures near volcanic vents in the deep ocean, feeding on carbon dioxide and spitting out minute amounts of methane--the natural gas used for power and home heating. Venter and his colleagues sequenced the bug's genome in 1996. In a Venterian world scientists would identify and harvest the particular genes that let it do its tricks, insert multiple copies of those genes into a synbug to amplify its methane-making ability, and--presto!--create a biological energy source that's economically competitive and environmentally saintly.

Scientists' dismay with Venter's renegade behavior during the human-genome race is nothing compared to the controversy his current project is likely to kick up if it starts breeding synbugs. From the start Venter has tried to defuse criticism by seeking the blessing of the ethics establishment. A University of Pennsylvania bioethics panel signed off on his proposal to build a minimal genome in 1999--but not without caution, says Caplan, who sat on the panel. "We didn't see any reasons not to proceed," he says, "but we saw many things to pay attention to." Those included the work's potential biowarfare applications and the environmental risks posed by escaped bugs. The risks haven't changed, Caplan says; science is still sorely in need of rules governing synthetic biology. "Nothing that has happened with genetically modified plants and animals makes me feel secure that those rules are in place," he says. "The time to draft them is now, when research is just beginning."

Others believe Venter needs to be stopped. "He's trying to short-circuit millions of years of evolution and create his own version of a second genesis," says Jeremy Rifkin, the longtime technology gadfly. Rifkin worries that, whatever reassurances Venter may offer, he can't control the consequences when--not if--a bug makes it into the environment or into criminal hands. "It's the height of hubris. It's irresponsible. He can't tell you that it's going to be safe."

Even so, the U.S. government, despite its post-9/11 cautiousness, has seen no problem in helping fund Venter's work. The scientific paper documenting how he built a virus from scratch received extensive security vetting by the administration, which gave it a pass.

Venter pooh-poohs the fears. "There's very little motivation for somebody to resort to complex molecular biology" to commit biological terrorism, he says. A terrorist "can get infectious agents from any hospital.... He can easily get anthrax from any farm." As for the environmental worries, he claims that any bug he synthesizes in the lab will be so delicate that it would never survive in the wild.

As this story is being written, Venter is back aboard Sorcerer II. This time he's in the South Pacific, bound for the Galapagos. There, where Charles Darwin unlocked the secrets of the natural world, Venter is seeking raw material for his brave new one. He's collecting seawater samples and will come back to sequence the millions of organisms he retrieves. Very likely they'll yield new genes to help soup up his model A--if he ever gets it running.