Pfizer restarted its shelved 2003 SARS program to build the Paxlovid pill in a year

In the early pandemic, attention and money went to vaccines — the US announced Operation Warp Speed in May 2020 — but Pfizer's research facility in Connecticut quietly restarted ideas the company had developed during the 2003 SARS outbreak. The approach was to block a well-understood part of the virus life cycle: a key protease that orchestrates how the virus copies itself. Find a chemical that sticks tightly to that protein and the virus stops replicating.

The first break was luck. None of the thousands of proteins in the human body shared the bit of molecular structure Pfizer planned to interfere with in SARS-CoV-2, so the drug could hit the virus without expected side effects. Pfizer chemist Dafydd Owen called it "the most solid biological target I have ever worked on." The company tested hundreds of chemicals in parallel, scaled up the most promising one, and ran animal studies even as the first vaccines were authorized in December 2020; human trials began in March 2021.

By fall 2021 a monitoring board stopped the human study because covid-19 patients on the drug weren't dying while those on the dummy drug were. A Pfizer test in unvaccinated volunteers showed the pill cut the chance of a serious case of covid by 89%. Charlotte Allerton, Pfizer's head of medicine design, believed it set a record as the fastest any drug company had moved from synthesizing a new chemical to proving it safely treats a disease. The Biden administration authorized the pill's sale on December 22, 2021, and the president publicly touted it.

The oral formulation was itself a decision: remdesivir, the first chemical approved in the US to treat covid-19, had to be infused intravenously five days in a row, which limited its reach. Pfizer's chemists tweaked their antiviral so a patient could swallow it — "You get a prescription, you go to the CVS, and that's it," as researcher Kris White put it. And because the protease rarely changes, laboratory tests suggested Paxlovid would work against any variant, possibly against all coronaviruses, even one still lurking in a bat cave.

Restart the shelved program aimed at the older coronavirus instead of starting from zero

The protease is highly conserved, so a drug aimed at it keeps working as the virus mutates around vaccines.

No human protein shares the target's molecular structure, so strong inhibition was not expected to cause major side effects.

Testing hundreds of chemicals in parallel and scaling the winner saved months against a sequential process.

An oral pill reaches patients through pharmacies, where remdesivir's five-day IV infusions tied treatment to hospitals.

Shelved R&D is stored option value: a program stopped when its market disappeared can be the fastest answer when a new one appears. A conserved target buys durability against mutation.

The Biden administration spent $5.3 billion in December 2021 to pre-purchase 10 million courses and doubled the order weeks later, though the 20 million courses would not all arrive until midyear, mostly too late for the omicron surge. Supply stayed short, the pivotal trial involved only about 2,000 people, and the pill had to be taken within five days of symptom onset. Pfizer's lab tests suggested it would work against all coronaviruses, raising the idea of stockpiling it against future pandemics.