At 7 A.M. on a midsummer day in Piracicaba, a city in the Brazilian state of São Paulo, Cecilia Kosmann sat in the back of a van surrounded by plastic take-out containers filled with genetically modified male Aedes aegypti mosquitoes. Every two minutes or so, she shook a container through a plastic funnel, releasing them into the cool outside air.

By the time the van finished its daily route through two neighborhoods here, she’d released about 250,000 of them.

These mosquitoes will spend their lives competing, copulating, and, because they are so numerous, overwhelming the population of wild males in the pursuit of females. Because of a genetic change to their DNA, they will live only four days—and their offspring won’t ever develop past the larval stage.

The insects were developed by Oxitec, a U.K. company that calls them “Friendly Aedes” and produces them at a facility located an hour away by car. Although the insects aren’t yet commercially available, the pilot program in Piracicaba has become a test case for whether GM insects can stop disease—and, if so, whether it will be at a cost cities can afford.

The project here began in April 2014, a year after an epidemic of dengue fever that caused more than 1.5 million cases in Brazil. So far, it is working: after 10 months of testing in two small neighborhoods, the number of dengue cases among 5,600 residents fell from 133 in a year to only one. The city’s mayor, Gabriel Ferrato, says he is seeking financing from the state to expand the program. “If I had the necessary resources right now, I’d adopt the method to cover all of Piracicaba,” he says.

Money isn’t easy to come by right now in Brazil, which is going through a major economic and political crisis. Its currency has plummeted, and an oil-money corruption scandal has reached the president’s office. Now it’s the epicenter of the Zika epidemic, too, and politicians are trading blame for that. There’s speculation that other countries could boycott the Summer Olympics in Rio. “We’re losing the battle against the mosquito in an ugly way,” Brazil’s health minister said in January.

But the problem is also turning the country into a hothouse for new technology. In addition to GM mosquitoes, there are experiments using mosquitoes infected with a bacterium, Wolbachia, that seems to prevent them from spreading diseases. Research on that idea has been supported with $40 million by the Bill & Melinda Gates Foundation, which hopes it could be extensively deployed in Brazil in the coming months. A still more futuristic approach, which would use gene editing to eradicate mosquitoes, remains years from actual use.

Last August Oxitec, which has also staged releases in Panama and the Cayman Islands, was acquired for $160 million by the U.S. conglomerate Intrexon, which owns a portfolio of transgenic organisms including salmon and apple trees. Now that Brazil is fighting Zika as well as dengue—both are spread by Aedes mosquitoes, as is the chikungunya virus—interest in the technology has surged. “The moment of crisis will pass, but we’re sure that our technology is here to stay,” says Glen Slade, director of Oxitec’s Brazil operation.

The GM insects are bred nearby in the city of Campinas, at a facility that can produce two million mosquitoes a week. In an all-white room, mosquitoes are mated and the resulting larvae divided by sex. Workers whisk at stray mosquitoes with electrified tennis rackets—the kind you see in novelty stores, but which have sold out in mosquito-obsessed Brazil.

Decades ago, Brazil nearly triumphed over Aedes by spraying insecticides and mobilizing people to clear standing water where the insects lay eggs. The aim then was to combat urban yellow fever, and by 1958, Brazilian officials had declared Aedes eradicated. But the mosquito returned, possibly on a foreign ship during the 1970s, finding ready breeding grounds amidst the chaotic expansion of Brazil’s cities and shantytowns.

“Even a tiny amount of water, in a candy wrapper tossed somewhere, can be a breeding ground for Aedes eggs,” says Margareth Capurro, a professor of biosciences at the University of São Paulo, who has worked with Oxitec. “We have to use all control methods available now—all imaginable weapons.”

The release of Oxitec’s mosquitoes follows years of research and promotion inside Brazil. Although they were approved in 2014 by Brazil’s biosafety agency, they can’t be sold commercially until the company wins certification from the country’s medical regulator.

The GM mosquitoes aren’t without critics. A conspiracy theory circulating on the Internet places them near the epicenter of the Zika outbreak, and the rumored connection, though baseless, is as hard to stamp out as the insects themselves. The involvement of a profit-making company only spurs the speculation. “GMOs are accepted when they are in the research phase, but when there is a company behind them, some people begin to view the commercialization of research as wrong,” says Capurro.

Oxitec’s strategy could end up being more expensive than some alternatives. That’s because for it to work, huge numbers of mosquitoes—greatly outnumbering the wild population—have to be released continually. For instance, Oxitec says that to protect about 5,600 people who live in the test area, it has been releasing between three and four million mosquitoes a month. Protecting all of Piracicaba, where Oxitec plans to build a new production site, would theoretically take three billion insects a year.

Oxitec has not said how much the mosquitoes cost, but officials in Piracicaba said they expected to eventually pay about 30 Brazilian reals, or $7.50 a year, per person protected. The bill for the city of 390,000 would come to about $2.7 million a year. That is approximately what the health department currently spends on sprays, larvicides, and costs like sick leave.

But city officials believe they can switch over and rely entirely on GM mosquitoes. “The reality is that with the success of the Friendly Aedes, the traditional methods can be abandoned,” says Carlos Eduardo Luccas Castro, a spokesman for the mayor’s office.

Although Oxitec technically can’t sell the mosquitoes yet in Brazil, Ferrato’s office said in January it had agreed to spend $800,000 over two years to expand the program to protect 60,000 people. In an online FAQ, Oxitec calls the payment a “contribution” to an “optimization” study.

Oxitec first developed the genetically modified mosquito in 2002, adapting the well-known strategy of releasing sterile males to compete for mates. Its mosquitoes aren’t sterile; instead, they’re engineered to vastly overproduce a protein that kills them unless they are fed the antibiotic tetracycline, as they are in the production center. Without the antidote, the released males quickly die off, as do their offspring.

Because of a second genetic modification—a fluorescent marker—the GM larvae glow when exposed to a special light, which allows field teams to count them in breeding grounds. Data collected by the city and Oxitec suggest the program has decreased the number of wild mosquitoes by more than 80 percent in the treated neighborhoods.

Pedro Mello, the city’s secretary of health, believes vaccines will be what finally controls the dengue and Zika viruses. But until that happens, he says, “we need to invest in technological methods of pest control.”

“Brazil is still using methods that have been employed since the beginning of the 20th century,” says Mello. “The GM mosquito goes to places where the others methods can’t reach—the dark, cool places behind furniture and under tables, inside people’s homes and their surroundings.”

Ferrato, the mayor, has taken a gradual approach to Oxitec’s bugs so far, letting people get used to the idea of GM insects buzzing around their houses. But he has bolder plans. With Oxitec opening a new mosquito factory in the city, he thinks Piracicaba may have hit on a new export industry. “We can become a supplier of mosquitoes to any other municipality in Brazil,” he hopes.