Last fall John Melo, the CEO of the synthetic-biology company Amyris, traveled to Amsterdam to visit one of his company’s largest investors. During the long weekend, the former oil executive says, he took a side trip to hang out in the city’s famous “coffee shops.”

He wanted to interview pot smokers.

“What made you do this, how often do you do this?” Melo says he asked.  “I really wanted to understand the psyche here. What are they looking for, and what do they get?”

The reason for the questions was that Amyris was about to get into cannabis, too. The company, which previously showed that it could use baker’s yeast to make jet fuel, a malaria drug, and no-calorie sweeteners, now believes it can genetically engineer the same organisms to manufacture cannabinoids, the active molecules in marijuana, more cheaply than by growing plants, which require a lot of water, space, and electricity.

But if this worked, he understood, cannabinoids could end up everywhere, in new forms, and reach new people. So Melo needed to resolve his own doubts. “I spent several years saying absolutely not,” he says, adding that his wife was “completely against” the idea.

His company, however, needed the business. Its biofuel venture had failed to be competitive, leaving a billion-dollar crater in its balance sheet. Now Melo wanted to know what a world awash in cannabinoids was like. “It was market research that led me to have moral research. Is there a moral question?” he says. “If we as a society consume [this] throughout the day to relax, what does that mean?”

Melo isn’t alone. As biotechnology embraces the legal cannabis industry, not a few of its executives are grappling with a dilemma: what will it mean if they make the drug much easier to get? 

So far, 33 US states have legalized medical marijuana, and 10 of these also permit recreational use. In October 2018, Canada permitted pot sales nationally. Health Canada expects that its citizens will consume 926,000 kilograms of cannabis this year, or one to two billion joints’ worth. Most will be consumed by daily pot smokers, who make up about 5% of Canada’s adult population, according to the agency.

Legalization has brought investors racing to stake their claim. Entrepreneurs in Denver erected so many greenhouses that the city’s electricity use surged noticeably. So it wasn’t long before genetic engineers realized they might be able to replace all those acres of plants. “In 2017, we started getting calls for scientists with experience in yeast, and I said, ‘What in the world ...?” recalls Kara Bradford, head of Viridian Staffing, a headhunter who specializes in finding agronomists and other specialists for the cannabis industry.

At first, Bradford says, the inquiries were highly confidential. Now, though, the race to make cannabinoids in microorganisms is out in the open. A dozen synthetic-biology companies, including Amyris, Ginkgo Bioworks, Hyasynth, Farmako, and Intrexon, say they are moving the cannabis genes into bacteria, oozing algae, or fermenting vats of yeast. This year, biologists at the University of California, Berkeley, inserted the critical steps in the cannabinoid recipe into yeast cells and reported the results in the journal Nature.

Cheap, pure supplies of cannabinoids could lead to new types of medicine and “wellness” products but also—if they’re added to things like soft drinks, cookies, and vaping fluids—to a big change in who can get high, when, and how. “My initial impulse was ‘Hey, what?” says Jason Kelly, CEO of Ginkgo Bioworks, a company in Boston that engineers yeast to make chemicals. “But it was clear they have a supply need, synthetic biology is a great fit, and it’s going to be a business.”

Other scientists admit that their work is cause for soul-searching. Biotech is an industry that ultimately wants to help people—and cannabis, like alcohol, doesn’t always do that. While fans tout benefits like pain relief, cannabis can also be habit-forming, and there is a link to psychosis, according to Health Canada.

“There is complete cognitive dissonance,” says one investor in Amyris, who didn’t want his name used. “Because ‘This is your brain on drugs,’ right? We all grew up with that.”

“There is complete cognitive dissonance, because ‘This is your brain on drugs,’ right? We all grew up with that.”

It remains to be seen whether biotech reactors can really make cannabinoids more cheaply than the fast-growing plant (it’s called “weed” for a reason). As well as THC, the psychoactive chemical that gets people high, and CBD, which has been approved in the US for treating certain kinds of epilepsy, marijuana contains another 100 or so far less concentrated cannabinoid molecules, which remain little studied because the plant has been illegal.

Jeff Ubersax, a biochemist who left Amyris to join a competitor, Demetrix, based on the Berkeley research, says his company thinks it can use biotechnology to manufacture larger amounts of these rare molecules and study what sorts of effects they have. “The really interesting thing with cannabinoids is [that] not as much research has been done on them, especially the rare ones—what do they do?” he says. “They are likely to be doing other really interesting and important things.”

The research remains tricky to carry out in the US because cannabis remains a schedule 1 drug at a federal level—that is, still illegal, like heroin. Biosynthesis research has to happen under unusual circumstances. Company research labs are equipped with security cameras, and yeast strains are kept in 750-pound safes inside cold rooms. Agents of the Drug Enforcement Administration certify the spaces.

“Cannabinoids are like the Uber of synthetic biology” says Karl Handelsman, founder of Codon Capital, an investment firm. “Is it legal or isn’t it legal? Is it a taxicab or not a cab?” He says he decided not to put money into the area out of concern investors could end up charged in a federal racketeering case: “It’s easy to say, ‘This has extra risk; I don’t want to do it.’” 

This March, following Melo’s research trip, Amyris announced that an investor group had agreed to pay as much as $300 million if it can figure out how to make cannabinoids for “large-scale use” in consumer brands. Melo says part of what changed his mind was learning over dinner one night that a close relative was using CBD oil for back pain. “Even at home, the whole resistance and perception of evil changed to good,” he says. He now also strongly believes CBD can treat opioid addiction (although there’s little evidence for that so far) and says that is the first molecule Amyris will attempt to make.

For a company like Amyris, bio-pot offers a new opportunity after the costly failure of its biofuel ambitions. “Hundreds of millions went into biofuels, but the best groups were making a dollar of fuel for eight dollars, so the field took a wrong turn,” says James Collins, a professor at MIT. Collins, considered one of the creators of synthetic biology, says the field has recently been seeking a new footing in health care, and so he questions whether making cannabinoids is a good idea. “I would be concerned about being labeled the marijuana company,” he says. “You have a coolness in some circles, but a stigma in others. It may detract from other serious business you try to do.”

Kelly, the Ginkgo CEO, says he feels that the issues of social acceptance have been mostly answered. In Massachusetts, voters agreed to legalize recreational marijuana in 2016. “The feeling was that the ship was sailing. The cultural winds are changing. We are in line with the majority of people in the US, that it’s a good alternative to other things people might be doing,” he says.

Ginkgo, an MIT spin-out, now appears to be trying to make the psychoactive ingredients in pot for use in global markets. In 2018, it announced a deal with Cronos, a fast-growing marijuana company in Toronto, which currently sells pre-rolled joints under names such as Spinach, a brand that describes itself as “geared towards a wide range of consumers that don’t take life too seriously.”

Cronos aims to become a “global cannabinoid company,” and for that it wants the molecules made cheaply and at large scale, not just in the form of plant flowers. Under the agreement, Cronos will pay Ginkgo $22 million to carry out research, plus another $100 million in stock, in tranches, if it is able to get yeast to make eight different cannabinoids at a cost of $1,000 a kilogram or less (a fraction of what it costs to extract the molecules from plants). Ginkgo will earn the single largest payment, 20% of the total, for demonstrating it can make THC.

In its own press statement on the deal, Ginkgo emphasized the possibility of developing new pharmaceuticals. But its partner, Cronos, says it also plans to mix cannabinoids and flavor molecules called terpenes together to create new, branded ways of getting intoxicated. “We’re using the plant as a blueprint,” Cronos CEO Michael Gorenstein said during a conference call with investors, “in order to learn and then create differentiated active ingredients” that will be formulated “to optimize bioavailability” and provided to consumers via “the appropriate delivery system.”

“[To] differentiate on psychoactive effects, so that we can actually build products that target different use occasions … that for us is the Holy Grail. That is where we see our brands heading,” Gorenstein added. “I am happy to report, this method of production is allowed; these are the same exact molecule as in the plant.”

Some public health researchers have been predicting that companies will pair legal cannabis with technology in novel ways. “There’s a whole lot you can do, in terms of going from the raw plant to the product that is actually consumed, that are much more sophisticated than what’s happened so far,” Stanton Glantz, a public health researcher and anti-tobacco campaigner, told Rolling Stone last year. 

While pot doesn’t hold the same dangers as opioids, whose recreational use has led to an epidemic of deaths, some health researchers worry about negative consequences if it’s made at new scales and delivered in new ways, like vaping. Already, teenagers are vaping nicotine in record numbers, and some are using the same types of electronic devices to vape THC in liquid form. A percentage of them will then develop what is called “problem cannabis use.”

The risks became clearer to some Ginkgo employees in December, when Cronos sold nearly half its shares to Altria, the cigarette maker, for $1.8 billion, with an option to take full control. Altria, headquartered in Richmond, Virginia, brings “significant expertise” to the cannabis business, according to Cronos, including its experience with vaping technology. Some people at Ginkgo are now not so comfortable with the cannabinoid project as they were before. That is because tobacco makers have a record of selling harmful products and lying about the science of addiction and cancer, as the US has proved in federal lawsuits. 

“Based on what I know, I don’t personally think this is a great leading image for the field,” says Megan Palmer, a public policy researcher at Stanford University who specializes in synthetic biology. “There are a lot of justifications that are possible; the question is whether you are accountable for the effects your product has in the world.”

Still, a big reason for legalization is to do away with the cannabis black market. (In the US, there have been hundreds of thousands of arrests each year for pot possession.) And you can’t achieve that policy aim without providing a legal product, notes Kevin Chen, head of Hyasynth, a Canadian genetic-engineering company that’s been working on cannabinoids. Chen told me he’s never smoked pot himself and that there are some products, like highly concentrated THC called “dabs”—smoking it produces intense effects—that he doesn’t want to be involved with.

“We don’t want to develop negative products,” he says. “But there is that opportunity out there which is not purely about helping people. Some of my employees and cofounders have had those discussions with their parents. Like, what are you doing?”