South of VaxGen's offices, the next freeway exit gives access to its powerhouse: Genentech. This is the world's front-runner in medical biotechnology, with seven licensed products, from human growth hormones to a clot-buster, Activase. Twenty years ago the company was all dreams and venture capital; its few staff snipping and splicing genes in a wasteland where shipyards had died. Today, their ranks of Mercs and BMWs surround 26 buildings in biology's Silicon Valley.
Stopping by from time to time are visitors from its master, the druggernaut Hoffman-La Roche. With twin headquarters in Basle and New Jersey, and sales last year of SWf24.7bn (œ10.2bn), this vitamins-to-Valium giant has the marketing muscle should AidsVax come on stream. At its own labs, Roche shuns the vaccine race, but with taxes pledged to line-and-jab Africa and Asia executives doodle in billions on the hope that Francis pulls it off.
When I flew to San Francisco to quiz Francis for this story, Berman was ecstatic, in jeans and a check shirt, over a new $1.4m vaccine facility. The experiment produces a torrent of clinic samples; each volunteer gives blood on 17 visits, and each sample is split for tests. Giant freezers were being installed to store bar-coded specimens. There could be 400,000 in all. There's also a $500,000 microbiology kit going in: DNA sequencers, PCR machines, centrifuges and the like. Soon he would direct 30 staff in 20 rooms. He was like a seven-year-old on Christmas Day.
The first thing that struck me was the push of the spending, irrespective of scientific achievements. Apart from all the investment so far, Genentech had a 10,000-litre fermenting tank, half full of New Jersey strain vaccine. Nobody wanted that, worth $1m, flushed away, much less the careers of its makers. The next thing I noted was the standard of safety imposed on the facility's construction. To handle a dangerous pathogen in California, the brown-and-yellow building, made from tipped-upright concrete slabs, was stamped with certificates and permits by the box load before the first plank was sawn. It's both earthquake- and microbe-proof. And its forests of copper pipes, air ducts and bio-filters were tested to tolerances few structures could endure.
But while regulations make sure that the building is safe, critics say that the product itself escapes much rigorous scrutiny. With vaccines, any problems often don't appear until mass-market use, and such is the head of steam building up behind Francis that sceptics think that if AidsVax doesn't join the annals of useless shots, it has the potential to join, say, a 1960s measles vaccine that made the disease in those infected worse.
What worries some scientists is that because AidsVax provokes antibodies to its own specific gp120 strains, there's a risk that it may actually suppress the immune system's ability to combat other strains. On this thinking (the principle is sometimes called "deceptive imprinting") even if the junkies were protected against the New Jersey and Chiang Mai strains, they might die more quickly if they get infected with one of the countless other mutations. "There's nothing new in this," Dr Heinz Kohler, who has led investigations at Kentucky University, said. "It's just common sense."
At Kansas University, researchers have found that monkeys injected with gp120 and then a hybrid kind of HIV had more of the virus in their blood later on than infected animals which weren't vaccinated. "The question is: will those people who are vaccinated progress to Aids more quickly if they become infected with HIV than those who were not vaccinated at all?" Prof McMichael at Oxford summarised. "We might not know the answer for 10 years."
No such problems were revealed in the preliminary tests, but despite the importance of long-term follow-up (recipients of the hepatitis B vaccine that Francis worked on in the early 1980s have been tracked for two decades), VaxGen no longer monitors what has happened to the people who received its product in the mid-1990s preliminary tests. Francis argues that it makes more sense to wait for the full-scale trial results.
This apparent loss of data is surprising to some, because history warns of the pitfalls of not being thorough. In 1955, just one month after a near-hysterical press conference in Michigan launched polio vaccine, reports poured in to the CDC of hundreds of children going down with the disease, induced by the shots themselves. President Dwight Eisenhower said that, because of the "great pressure to bring this out", scientists may have "short-cut a little bit".
AidsVax cannot give volunteers Aids, but there may be something even more terrifying than the anxiety that it might accelerate their disease if they are later infected with HIV. Some scientists think that, if it works at all, the product may have a dangerous effect on the evolution of HIV. Five years ago, Los Alamos scientists declared that there was "no simple answer" as to whether Aids could become contagious through coughs and sneezes - and other researchers argue that, in much the same way as a partial course of antibiotics can promote resistant bacteria, so a poorly-effective vaccine may promote more deadly and infectious strains.
This may sound like journalistic scare, but HIV's best-understood RNA cousin is influenza virus, which produces devastating mutations every 20 or 30 years. Hepatitis B virus, meanwhile, has already produced mutant strains accepted as being vaccine-induced. "When you use a vaccine, you are introducing another selective pressure," Dr Paul Ewald, professor of biology at Amherst College, Massachusetts, explained. "It could make the problem more damaging, or less damaging, depending on the antigen you use."
Researchers told me that, compared with the potential risks to volunteers, this doomsday scenario was "unlikely". But with agencies standing by to jab hundreds of millions of people, some wondered if, for our species' safety, "unlikely" was reassuring enough. "My personal view," Dr Art Ammann, president of the San Francisco-based Global Strategy for HIV Prevention, and a former AidsVax researcher, said, "is that we could face a global nightmare."