Flagship Pioneering’s offices sit right on the water in the Kendall Square neighborhood of Cambridge, Mass., with floor-to-ceiling windows overlooking a breathtaking view of the Charles River as it gives way to Boston’s downtown skyscrapers. On a sunny spring day, the view from here can make the misery of a Bostonian winter feel worthwhile.

For many of the folks working behind these windows, history can be divided into two parts: Life Before Moderna and Life After. Flagship Pioneering is the biotech venture capital firm best-known for funding Moderna, a startup that went from also-ran to household name after it created a highly effective COVID-19 vaccine based on its mRNA technology.

Moderna has sold and distributed billions of doses of that vaccine to date, turning it into a Fortune 500 company and a stock market darling. The company brought in $18 billion in revenue in 2022, and its success transformed multiple founders into billionaires. But that windfall also turned Flagship itself into by far the most famous firm in the normally low-key club of biotech VCs. Noubar Afeyan, Flagship’s founder and CEO (and chairman of Moderna), has seen his face splashed across magazine covers, while profits from Moderna’s runaway success filled Flagship’s bank accounts and gave its investors global bragging rights.

Even before the COVID pandemic, Flagship had enjoyed a string of notable if more modest successes and had earned respect in the field for its unusual approach: It invests almost exclusively in startups that it incubates in-house. Today, it’s in the enviable position of trying to figure out what its next big hit might be. You could think of it as planning Life After Life After Moderna—but of course, the long shadow of Moderna’s success looms large in the consciousness of Flagship’s employees.

“Our goal is essentially to create the next Moderna,” a Flagship employee tells me during a visit to the Kendall Square offices. (The employee asked not to be identified because we were discussing confidential research.) “That’s what we all want to do. That’s why we’re here.”

Flagship grabbed headlines in July with the announcement of a new $100 million partnership with pharma giant Pfizer. Both companies will invest $50 million upfront to explore 10 drug programs across Flagship’s full portfolio of companies. Flagship will be eligible to receive up to $700 million in payments for each successful program, meaning the program is potentially worth up to $7 billion in total. That partnership—ironically, a tie-up with Moderna’s top COVID vaccine rival—may someday produce Flagship’s next blockbuster. Or that blockbuster could come from one of its many other portfolio companies—or it may never come at all.

Over the course of a few weeks this spring and summer, Fortune spoke with Flagship’s leaders to learn about its innovation process and get a sense of how the company’s culture is evolving as it aims to build on its successes.

Founded in 2000 by Afeyan, Flagship Pioneering is part VC firm, part startup accelerator, part research lab. Like other VCs, it provides capital to early-stage biotech startups with the aim of recouping that investment many times over in the future. But it’s also a startup accelerator because it helps founders turn new ideas into companies. And it’s a research lab because it endeavors to discover and develop those new ideas with its own crack team of scientists—many of whom it recruits from nearby Harvard and MIT, and many of whom work for both Flagship and the startups as companies launch. It is essentially a venture capital firm staffed almost entirely by scientists.

But finding the next Moderna will be complicated to say the least. For starters, the unique market conditions created by COVID gave Moderna a boost that few other companies could hope to enjoy. The other companies in Flagship’s pipeline are still in early, high-risk stages where a bad clinical trial could drive them into oblivion. And of the two companies Flagship itself identifies as most promising, one—gene-editing startup Tessera—has entangled itself in a controversy over alleged intellectual copycatting. Those allegations have fanned long-simmering complaints from Flagship’s detractors—most of them scientists and executives at competing biotech firms—that the firm is self-aggrandizing and opportunistic, and that its reputation for being uniquely innovative isn’t merited.

Flagship leaders shrug off such criticism, saying that their firm’s combination of ambitious research and general disdain for the status quo of entrepreneurship make it an easy target for the envious. And Afeyan makes the case that Flagship’s approach to science, in addition to its individual discoveries, makes it stand apart. 

“Flagship is, above all else, an experiment in the possibility that the mechanism of creating companies could become a profession,” Afeyan tells me, as he reclines languidly in a desk chair at a conference room table.

So far, that mechanism has meant creating companies quickly, abandoning many of them just as quickly, and counting on the successes to outweigh the gaffes.

Building companies with ‘paranoid optimism’

Flagship was originally called NewCoGen, which stood for New Company Generation, before changing its name to Flagship Ventures in 2002, and then again to Flagship Pioneering in 2016. The firm’s various name changes are consistent with its evolution to the VC and R&D hybrid it is now. In its early days Flagship also invested in startups that it didn’t incubate and that had little to do with biotech, including a marketing tech startup and a compliance software company. 

Today, the focus is squarely on biotech, and on prioritizing big swings over incremental improvements. Flagship’s model of venture capital, which it calls “venture creation,” starts with a kernel of an idea, often on the brink between kooky and groundbreaking; it says it pursues only those with the potential to be revolutionary. “Leadership in our realm is systematically trying to avoid dedicating our time and effort towards something that doesn’t have an incredibly high chance of changing the world,” says Flagship partner Geoffrey von Maltzahn, explaining what is considered a Flagship-worthy idea.

Flagship usually researches between 80 and 100 such ideas a year. Between 10 to 15 of those ideas will eventually yield some promise of scientific viability and become the basis for early-stage startups. The notable caveat, though, is that the roughly $1.5 million in pre-seed money that each startup receives comes exclusively from Flagship. If the science pans out after preliminary lab experiments, the startups start to scale in the conventional sense—raising money from other investors after a Series B round, determining a product pipeline, and hiring employees.

Currently, Flagship has 45 portfolio companies to go along with 38 exits, according to a company spokesperson. Its biggest successes other than Moderna include Receptos, which was developing a multiple sclerosis drug and was acquired by Celgene for $7.2 billion in July 2015 (Flagship had sold all its shares prior to the acquisition); and Acceleron Pharma, which had produced a drug to treat rare blood disorders and which Merck bought for $11.5 billion in September 2021. The firm has also been attached to at least 30 comparatively smaller deals—under a billion dollars—since 2000. In June, for example, it sold portfolio company Sigilon Therapeutics to Eli Lilly in a deal worth up to $310 million. Flagship has also been involved in 30 IPOs.

As with all venture capital endeavors, there have also been plenty of washouts. Biotech startups don’t derive their value from medicines currently in the market, but rather from their potential to create new ones should their scientific research bear out—and when it doesn’t, investors usually draw the curtain on the company. Twenty of Flagship’s portfolio companies have gone out of business since 2003.

Like most VC firms, Flagship doesn’t publicly report its overall returns on capital. But its track record, with a relatively high ratio of wins to losses, not to mention the size of those wins, makes it clear that it has rewarded investors well, and Afeyan carries himself like someone riding a winning streak.

Afeyan, 61, is a hulking man who looks more like an aging linebacker than a scientist.

Born to Armenian parents in Beirut, Afeyan moved with his family to Canada in 1975 to flee Lebanon’s civil war. In 1987, Afeyan earned a Ph.D. from MIT in the then-burgeoning field of biochemical engineering. Ten years later, at just 35 years old, he sold his first company, Perseptive Biosystems, which made laboratory equipment, to Perkin-Elmer Corp., in a stock swap deal worth $360 million.

He has knowing eyes and a fondness for coining his own catchphrases. He describes Flagship’s mentality as one of “paranoid optimism,” and he regularly coaches the industry veterans he hires to run Flagship’s portfolio companies that “it’s hard to expect unreasonable outcomes by being reasonable.” He’s a charismatic speaker, possessing a disarming self-awareness. At one point we discuss medication for treating obesity, which, he casually remarks, “obviously I suffer from.”

Afeyan also has a habit of saying ambitious—or outlandish—things with such nonchalance that they suddenly seem straightforward and obvious. When I ask him which, if any, of Flagship’s portfolio companies have the possibility to be the next Moderna, he responds: “The reality is every one of our companies that are operating today actually think they’re going to be bigger than Moderna. That is the conceit they all are going to have.

“And by the way,” he adds, “Moderna used to think it was going to be bigger than everything else.” Afeyan says he’d originally envisioned Moderna as a $100 billion company. (Its market capitalization topped $180 billion in 2021, after its vaccines rolled out; in early August, it stood at $41 billion.)

What justifies the bullishness of Afeyan and his Flagship teammates, in their view, is Flagship’s preference for developing biotech platforms—rather than one-off products—that if successful could be applicable to all manner of medicines. Moderna’s mRNA vaccine technology, which could, in theory, be applied to multiple diseases, is the quintessential Flagship platform. It had conducted successful Phase I trials for nine other mRNA vaccines before pivoting to COVID-19 in early 2020. 

When I ask him to narrow down his company’s current contenders, Afeyan says he’s thinking of portfolio companies with “interesting critical mass from an execution standpoint.” He likened watching the advanced stage of those companies to catching a close baseball game in the fifth inning, when the end of the game is closer and the stakes higher, as opposed to the first inning when there’s still the whole game to be played. (Afeyan loves a good sports metaphor, and sports in general. At one point, he interrupted our interview to talk about attending the 2022 World Cup in Qatar.)

Two of the companies Afeyan says are furthest along and perhaps closest to the lofty goal of being the “next Moderna” are Generate Biomedicines, a startup using machine learning to engineer new proteins in attempts to develop treatments for cancer, infectious disease, and immune disease; and gene editing company Tessera Therapeutics. Neither has any drugs or products on the market yet. But both have far-reaching goals.

Now valued at $1.45 billion, according to PitchBook, Generate is focused not simply on producing new drugs, but on finding an entirely new way to create them. Generate’s platform is premised on using machine learning to create a generative A.I. tool that analyzes the relationship between protein structures and amino acid sequences of hundreds of thousands of proteins to predict which one is best suited to be the key ingredient in a given new medicine. “If this hypothesis is right,” says Generate cofounder, chief strategy officer, and chief innovation officer Molly Gibson, “which we’re seeing more and more evidence of across the industry, then this is the way all protein therapeutics are going to be done in the future.”

Traditionally, scientists attempting the same exercise have had to use a much slower test called a high-throughput screen that requires repeated trial and error to design the right protein. Machine learning eliminates much of that guesswork, making the process roughly 100-fold faster, says David Baker, head of the Institute for Protein Design at the University of Washington. “It’s almost like DALL-E,” says Baker, referencing OpenAI’s text-to-image tool. “To design a protein…you give it minimal inputs, and then basically the A.I. figures out a lot of the details.”

Generate CEO Mike Nally, who was chief marketing officer of Merck’s human health division prior to joining Flagship in March 2021, says his company has 15 drugs in preclinical trials, the stage of the FDA approval process before a new medication can be tested in humans. The company has also forged some promising partnerships. In January 2022, Generate signed a research agreement with Amgen that includes $50 million upfront and up to $1.9 billion in total payments, not including future royalties should the partnership yield any drugs. It also inked a co-development deal this April with the University of Texas’s MD Anderson Cancer Center to collaborate on five cancer drugs.

“Drug discovery has historically been an artisanal craft, down to an individual genius to come up with a breakthrough idea,” Nally says. “What computers have in terms of an advantage over that traditional method is computers are inherently scalable. It’s hard to scale genius; there was [only] one Leonardo da Vinci.”

Plenty of failures alongside the hits

For every Generate that gets the Afeyan stamp of approval, there are scores of ideas that don’t become startups, dozens of startups that don’t become companies, and some companies that fail. In the past three years, four Flagship companies have gone belly-up.

Kaleido BioSciences, a company specializing in microbiome therapeutics, shuttered in April 2022, just over three years after its $75 million IPO. It failed to find a buyer after receiving a warning letter from the FDA over “objectionable conditions” in its clinical trials. A year before that, Flagship’s Ohana BioSciences closed its doors just 18 months after launch. Codiak BioSciences, another publicly traded Flagship company, filed for Chapter 11 bankruptcy protection in March. Bob Nelsen, cofounder and managing director of VC firm ARCH Venture Partners, which holds multiple co-investments with Flagship, says his firm lost between $30 million and $40 million as a result of the bankruptcy. Flagship’s biggest miss of the four was Rubius Therapeutics. Rubius’s July 2018 initial public offering of nearly $2 billion was the largest ever for a biotech firm at the time, according to biotech trade publication STAT.

Only Ohana failed because of a “weakness in the platform,” Afeyan says. The others were simply subject to the ups and downs of trying to secure venture capital funding. Rubius, Afeyan admits, may have also had some uncertainties with its underlying technology. The company’s efforts to convert healthy red blood cells into drugs failed to yield positive results for treatment of the rare genetic disease phenylketonuria (commonly known as PKU); its efforts whiffed again once the company pivoted to cancer treatments.

Afeyan takes these failures surprisingly well—another hallmark of the “Flagship way.” Flagship coaches its employees (particularly the younger ones actually cooking up research ideas that might one day become companies) that setbacks, even major ones, are necessary in order to reach the outsize success of a Moderna 2.0. “A lot of people who want to succeed are willing to succeed less, but fail even lesser,” he says.

The nature of venture capital means that Flagship can afford, even expect, some misses in its portfolio. But Flagship now has a much higher public profile than your average biotech VC firm. And since it claims to only invest in revolutionary science, the failures of its less-promising companies seem more jarring, because they were originally billed as such breakthroughs.

Some critics think that Flagship’s leaders set themselves up for second-guessing and schadenfreude by publicly touting the innovativeness of their process—which even became a Harvard Business School case study. “I think they’re full of shit when they say that, but that’s okay, they can drink their own Kool-Aid,” says François Vigneault, founder and CEO of Shape Therapeutics, also a protein engineering startup. But Vigneault concedes that some individuals at Flagship are talented at first-principles thinking, citing von Maltzahn by name. (In June, Seres Therapeutics, a Flagship company cofounded by von Maltzahn, was named to Time’s 100 Most Influential Companies list.)

Flagship and Afeyan point to the success of Moderna’s mRNA vaccine as the ultimate validation of their approach. The novelty behind the mRNA vaccine is that it co-opts the body’s own cells to manufacture an antibody, rather than activating antibodies by injecting a weakened form of the virus into a patient. It was this novelty that elevated it for incubation status at Flagship, way back in 2010 when the company launched, and that fueled its more recent achievement.

But the recent triumph of the vaccine obscures the fact that it took more than a decade of research to perfect the technology. Moderna went public in 2018; as recently as 2019, it had annual revenues of only $60.2 million, according to SEC filings, entirely from research grants and external R&D funding it received through partnerships with Big Pharma houses including Merck and AstraZeneca.

In 2020, revenue jumped to $803.4 million, as the company secured tax dollars and grant money to accelerate production of its newly developed COVID-19 mRNA vaccine. Only in 2021, once vaccines started to become widely available, did revenues leap to $18.5 billion, leading Moderna to profitability. (It earned $12.2 billion in profits in 2021.) Sen. Elizabeth Warren (D-Mass.) has estimated that Moderna received nearly $10 billion in federal government funding as it scaled up manufacturing of its COVID-19 vaccine.

“They had a set of tools, and when COVID happened, they capitalized and actually did really, really well,” says Vigneault. “But I would argue that everything else Moderna was set up for has failed.”

Flagship’s fans see such carping as sour grapes. “What made it even more poetic is how many people shit on [Moderna] all the way through its existence, until it saved the world,” says ARCH Venture’s Nelsen, with a laugh, speaking on a Zoom call from Palm Springs as he sported a pair of turquoise Vuarnet sunglasses. “I’m sure from a Flagship perspective it was sweet to have that kind of vindication.”

A copycat controversy

Tessera, the gene-editing company Afeyan cites as a promising upstart in Flagship’s stables, has been operating as an independent firm since 2018. It’s currently valued at $1.7 billion with $580 million in venture capital funding, according to PitchBook. Tessera’s technology is based on what the company calls “gene writers” that aim to cure genetic diseases by fixing errors in the human genome. At a trade conference in May, Tessera presented updates on its sickle cell disease research and data from nonhuman primate trials for PKU.

Prior to the conference, however, Tessera had become the subject of troubling accusations. Critics accuse Tessera Therapeutics of essentially copying “prime editing,” a breakthrough in gene editing discovered by Broad Institute scientists David Liu and Andrew Anzalone and published in a seminal paper in Nature in 2019. (The Broad Institute did not respond to multiple requests for comment.) Prime editing’s promise lies in its ability to allow researchers to insert, swap, or edit more than one base pair of the human genome at once at a specific location. If it proves applicable in human patients, prime editing could offer a means of curing diseases including sickle cell anemia, cystic fibrosis, and genetic blindness.

As of now, prime editing can still edit only about 40 base pairs at a time among the billions of pairs in the human genome. For that reason, some scientists, including those at Tessera, have been trying to crack the code for a version of gene editing known as retrotransposition that would allow for the editing of thousands of base pairs at once.

As of now, Tessera has claimed some advances, but it hasn’t published any peer-reviewed work showing its scientists have succeeded in solving retrotransposition and declined to share progress when asked by Fortune. Multiple sources, including academics and other biotech executives, told Fortune that they believe Tessera has simply pivoted to using prime editing technologies in its research without admitting as much publicly—and, more important, without licensing the technology from the Broad Institute, which holds its patent rights.

“Tessera is using prime editing,” Dr. Kiran Musunuru, director of the genetic and epigenetic origins of disease program at the University of Pennsylvania, told Fortune in an email. “Although their public presentations and marketing obscure that fact, their patent applications published in March make it clear.” Musunuru points to conspicuously similar diagrams in a Tessera patent filed in 2022 and a diagram in the 2019 Nature paper from the Liu lab. “[Tessera] even use[s] some of David Liu’s group’s terminology in the description of their work.”

When presented with these concerns, von Maltzahn, who is also founding CEO of Tessera, pointed to a patent application Flagship filed in August 2018 for target prime reverse transcription. That technology, he says, is a broader gene editing technology of which prime editing is a subset. He is careful to point out that neither he nor his partner Jacob Rubens discovered prime editing. When asked how the two processes were different, von Maltzahn replied: “We use distinct components and distinct systems. We haven’t described what those are,” attributing the vagueness on that point to the need to maintain a competitive advantage.

Tessera is under no obligation to share information it prefers to keep secret from its competitors. Von Maltzahn called the questions over Tessera’s use of prime editing to be “a largely academic debate” over terminology used to define which type of processes fall under the umbrella of the technology covered in Flagship’s 2018 patent filing. “’Is this prime editing?’ is actually not an IP question at all,” von Maltzahn says. Instead, he referred to it as a debate over “choice of lexicon to categorize things.”

Tessera isn’t the only Flagship company whose overlap with other companies’ technology has drawn attention. In a STAT story published in June about Flagship portfolio company Laronde, former Laronde employees said that the “endless RNA” central to Laronde’s methodology is essentially the same as the circular RNA technology in development at rivals Orna Therapeutics and Orbital Therapeutics. When asked for comment, Flagship pointed Fortune to statements provided for the STAT article, in which a spokesperson said there were differences in the “underlying sequences and elements” of the two RNA technologies.

The kind of behavior that Tessera has been accused of falls into the category of less-than-ethical, but not illegal—at least not yet. The Hatch-Waxman Act of 1984 allows companies to use patent-protected intellectual property in developing a therapeutic, explains Don Mizerk, a partner at law firm Husch Blackwell who has litigated patent disputes around well-known drugs like Adderall and Wellbutrin. The use of the IP only becomes copyright infringement if and when the drug gets commercialized and sold to the public.

“In the pharmaceutical space, you can research until the cows come home, and nobody can try to tag you with any infringing activity,” Mizerk explains.

While copycatting like that alleged at Tessera is relatively rare, disputes over who exactly should receive credit for what are common in the scientific world. “You don’t know who's first until you fight about who’s first,” Mizerk says. In biotech, where the business and scientific worlds intersect, companies might violate scientific norms in using someone else’s work without proper attribution, under the assumption that if they do get a product to market, they can make up for their behavior with a royalty agreement or another legal settlement.

“In my experience,” Mizerk says, “what usually happens in these cases is that if somebody is first, but they need the other person’s IP in order to actually market the product, then they generally strike some kind of business deal because they need each other to survive.”

Both Mizerk and Musunuru, the genetic-disease expert, point out that if Tessera were to be the first to develop a viable gene-editing therapy using prime editing, the company and the public would stand to profit enormously—and no court would want to bar Tessera from bringing the therapy to market because of an IP dispute.

Daniel Getts, CEO and cofounder of Myeloid Therapeutics, also a gene-editing startup, says that he sees the real problem with the alleged copycatting as being a waste of research funding and talent. “How much money is squandered in the meantime?” he asks. “How much real innovation will suffer simply because the $500 million or whatever could have been deployed to something real?”

Disputes as speed bumps

Where exactly Flagship ends and its portfolio companies begin is a question the new Pfizer deal only underscores. The answer is sometimes nebulous. For example, the CEOs of both Generate and Tessera are Flagship partners. Virtually every portfolio company has at least one board member who is a Flagship employee—von Maltzahn and Afeyan are both on Tessera’s board—and every company is, of course, birthed directly from Flagship itself.

In any case, there’s one factor that Flagship and its fleet of companies have in common: They’re relatively unfazed by criticism. In fact, the firm encourages its employees to embrace it and consider it a natural consequence of trying to score Moderna-size successes. When you’re in that mindset, disputes like the Tessera controversy can seem like speed bumps. As von Maltzahn put it, “We’re just way more focused on ‘How do we make great medicines?’ than ‘What should the terminology be?’”

Correction: An earlier version of this story said that Laronde, a Flagship portfolio company, had been accused of appropriating other companies' intellectual property. In fact, while former Laronde employees have said that the technology is essentially identical to that of some rivals, no one has openly accused Laronde of copying or otherwise appropriating it.

This story was originally featured on Fortune.com

More from Fortune: 
5 side hustles where you may earn over $20,000 per year—all while working from home
Looking to make extra cash? This CD has a 5.15% APY right now
Buying a house? Here's how much to save
This is how much money you need to earn annually to comfortably buy a $600,000 home