One Pandemic, Two Bets, Opposite Outcomes

What Moderna and Johnson & Johnson's COVID vaccine programs actually teach venture investors about science, risk, and luck

Prepared by Richstorm.co

Key Takeaways

• Flagship backed Moderna's mRNA platform in 2010 on the strength of a single peer-reviewed paper — no patent, no clinical data.

• J&J's COVID vaccine used a far more mature, clinically precedented platform — and was withdrawn after disclosed losses exceeding $1.5 billion.

• AstraZeneca's vaccine, built on a different adenovirus design specifically to dodge a known weakness, was withdrawn for the same root cause as J&J's.

• Even Flagship's own portfolio shows roughly one Moderna-scale outcome in about 148 companies — under 1%.

• The lesson for investors isn't which platform to back — it's that verified science narrows risk without removing it, and some risks are undiscoverable until mass scale.

Two Bets, One Pandemic

In 2010, Flagship Pioneering backed an idea with no name, no employees, and no proven product. The entire thesis rested on a single peer-reviewed finding: Harvard stem-cell biologist Derrick Rossi had shown that modified mRNA could reprogram human cells without triggering the immune response that had previously made the approach unworkable. That company became Moderna.

A decade later, a once-in-a-century pandemic gave that platform its proving ground — and gave a very different kind of company the same opportunity. Johnson & Johnson, one of the largest and most experienced pharmaceutical companies in the world, built its own COVID-19 vaccine on a platform with a far longer track record: an adenovirus vector technology already used in an approved Ebola vaccine.

One of these bets became one of the most consequential biotech outcomes in a generation. The other was withdrawn from the market, with disclosed exit costs exceeding $1.5 billion. Understanding why is a sharper lesson in venture risk than either story told alone.

What Each Side Actually Had — and Didn't Have

At the moment Flagship committed capital, Moderna's platform had real, checkable evidence behind it: Rossi's findings were published in Cell Stem Cell, a high-rigor journal, and the result was independently corroborated by a separate research group's earlier, related finding. What it did not have was a granted patent, clinical trial data, or any proof the platform could scale into a manufacturable medicine. The verified piece was the science. The unverified piece was everything about turning that science into a company.

J&J's and AstraZeneca's adenovirus vector platforms sat in the opposite position. The underlying technology was mature and well-precedented — J&J's own platform had already produced an approved Ebola vaccine, and adenoviral vectors more broadly had been studied in over 190,000 prior vaccine recipients before COVID. By any conventional measure of technical risk, this was the safer starting point.

That gap in starting risk is exactly what makes the outcome instructive.

The Outcomes, By the Numbers

Why the More Established Platform Still Lost the Race

The adenovirus vector platform that J&J and AstraZeneca used was not a flawed approach. It had already proven itself in an approved Ebola vaccine, and published reviews of its COVID-19 use describe it as one of the fastest, cheapest, and most effective ways to respond to a pandemic at scale, with a benefit-risk profile that remained favorable throughout. The platform did, however, carry one rare but serious side effect that the mRNA platform did not share: in a very small number of recipients — roughly one in every 26,500 to 127,000 doses — the modified virus used to deliver the vaccine triggered an unusual immune reaction that could lead to severe, sometimes fatal blood clotting. mRNA vaccines, which use a completely different delivery method with no virus particle involved, did not carry this same risk.

Notably, AstraZeneca had specifically chosen a virus from chimpanzees rather than humans, to engineer around a different known weakness — pre-existing human immunity to common cold-causing viruses. That was a genuinely sophisticated design choice, and it worked as intended. It did not, however, prevent the rare clotting risk, because that vulnerability traces to a feature shared across this category of virus-carrier vaccines generally — not to either company's specific engineering decisions, and not to a flaw unique to either vaccine.

mRNA vaccines don't introduce a viral particle into the body at all — the genetic instructions are delivered through lipid nanoparticles, a fundamentally different physical mechanism with no adenovirus structure present to trigger this specific pathway.

Just as important: this risk could not have been caught in advance. Phase 3 trials enrolling tens of thousands of participants are simply too small to detect a side effect occurring in roughly two to three people per 100,000 vaccinated. It only became visible once tens of millions of doses had been administered. No amount of additional diligence available in 2020 would have surfaced it sooner.

Even Flagship Mostly Loses

It's worth resisting the temptation to read Moderna's success as proof that Flagship's science-first model works. Flagship has backed roughly 148 companies. Three have reached unicorn status. One — Moderna — reached the scale and significance that defines this story. That's a hit rate under 1% for a true outlier outcome, from a firm built specifically around originating and evaluating scientific theses internally, with access to validators as credentialed as MIT's Robert Langer.

Even Noubar Afeyan, Flagship's founder, described vaccine approval itself as "a low-probability event" — a notable admission from inside the firm that made the bet. Moderna's outcome also depended on an unforeseeable global event: without a pandemic creating sudden, massive demand and compressed regulatory timelines, the company's pre-2020 pipeline was still a portfolio of early- and mid-stage candidates a full decade after founding, with no approved product.

What This Means for Evaluating the Next Bet

Put the two stories together and a few practical conclusions hold up, supported by what actually happened rather than by which story is more satisfying to tell:

• A more de-risked starting position didn't win. Moderna was the less-verified bet by every conventional measure — no patent, no clinical data, a single peer-reviewed paper. J&J and AstraZeneca had the more mature, more clinically precedented platforms. The less-verified bet succeeded; the more-verified ones failed.

• Platform maturity is not the same as safety. The more clinically precedented, more extensively studied platform carried the undiscovered risk. The newer, less-proven platform didn't.

• Some risks are structurally invisible at decision time, even when the underlying platform is sound. The clotting risk did not reflect a flawed technology — the platform had a strong track record and a favorable benefit-risk profile overall. It simply carried one rare, severe risk that no trial small enough to be run responsibly could have detected, and that a different platform happened not to share.

• Diversification is doing more work than selection. A scientifically embedded investor with unusually strong access still produced a sub-1% rate of true outlier outcomes. The portfolio, not the individual pick, is what survives that math.

None of this argues against scientific literacy in venture investing — understanding what a peer-reviewed finding does and doesn't prove, what a patent does and doesn't protect, and what a platform's structural risks might be is a genuine, defensible edge. What the Moderna and J&J comparison shows is narrower and more useful: correctly identifying good science did not determine which bet won. Moderna's science was real, but so was J&J's — the outcome was decided downstream, by a platform-level risk invisible at the time of decision, and by timing none of these investors controlled.