In March, 2020, researchers at Emory University published a paper about a molecule called NHC/EIDD-2801. At the time, there were no treatments available for the coronavirus. But NHC/EIDD-2801, the researchers wrote, possessed “potency against multiple coronaviruses,” and could become “an effective antiviral against SARS-CoV-2.” A few days later, Emory licensed the molecule to Ridgeback Biotherapeutics, a Miami-based biotechnology company which had previously developed a monoclonal antibody for Ebola. Ridgeback partnered with the pharmaceutical giant Merck to accelerate its development.

The Emory researchers named their drug molnupiravir, after Mjölnir—the hammer of Thor. It turns out that this was not hyperbole. Last month, Merck and Ridgeback announced that molnupiravir could reduce by half the chances that a person infected by the coronavirus would need to be hospitalized. The drug was so overwhelmingly effective that an independent committee asked the researchers to stop their Phase III trial early—it would have been unethical to continue giving participants placebos. None of the nearly four hundred patients who received molnupiravir in the trial went on to die, and the drug had no major side effects. On November 4th, the U.K. became the first country to approve molnupiravir; many observers expect that an emergency-use authorization will come from the U.S. Food and Drug Administration in December.

Oral antivirals like molnupiravir could transform the treatment of COVID-19, and of the pandemic more generally. Currently, treatments aimed at fighting COVID—mainly monoclonal antibodies and antiviral drugs like remdesivir—are given through infusion or injection, usually in clinics or hospitals. By the time people manage to arrange a visit, they are often too sick to receive much benefit. Molnupiravir, however, is a little orange pill. A person might wake up, feel unwell, get a rapid COVID test, and head to the pharmacy around the corner to pick up a pack. A full course, which needs to start within five days of the appearance of symptoms, consists of forty pills—four capsules taken twice a day, for five days. Merck is now testing whether molnupiravir can prevent not just hospitalization after infection but also infection after exposure. If that’s the case, then the drug might be taken prophylactically—you could get a prescription when someone in your household tests positive, even if you haven’t.

Molnupiravir is—and is likely to remain—effective against all the major coronavirus variants. In fact, at least in the lab, it works against any number of RNA viruses besides SARS-CoV-2, including Ebola, hepatitis C, R.S.V., and norovirus. Instead of targeting the coronavirus’s spike protein, as vaccine-generated antibodies do, molnupiravir attacks the virus’s basic replication machinery. The spike protein mutates over time, but the replication machinery is mostly set in stone, and compromising that would make it hard for the virus to evolve resistance. Once it’s inside the body, molnupiravir breaks down into a molecule called NHC. As my colleague Matthew Hutson explained, in a piece about antiviral drugs published last year, NHC is similar to cytosine, one of the four “bases” from which viral RNA is constructed; when the coronavirus’s RNA begins to copy itself, it slips into cytosine’s spot, in a kind of “Freaky Friday” swap. The molecule evades the virus’s genetic proofreading mechanisms and wreaks havoc, pairing with other bases, introducing a bevy of errors, and ultimately crashing the system.

A drug that’s so good at messing with viral RNA has led some to ask whether it messes with human DNA, too. (Merck’s trial excluded pregnant and breast-feeding women, and women of childbearing age had to be on contraceptives.) This is a long-standing concern about antiviral drugs that introduce genomic errors. A recent study suggests that molnupiravir, taken at high doses and for extended periods, can, in fact, introduce mutations into DNA. But, as the biochemist Derek Lowe noted, in a blog post for Science, these findings probably don’t apply directly to the real-world use of molnupiravir in COVID patients. The study was conducted in cells, not live animals or humans. The cells were exposed to the drug for more than a month; even at the highest doses, it caused fewer mutations than were created by a brief exposure to ultraviolet light. Meanwhile, Merck has run a battery of tests—both in the lab and in animal models—and found no evidence that molnupiravir causes problematic mutations at the dose and duration at which it will be prescribed.

With winter approaching, America is entering another precarious moment in the pandemic. Coronavirus cases have spiked in many European countries—including some with higher vaccination rates than the U.S.—and some American hospitals are already starting to buckle under the weight of a new wave. Nearly fifty thousand Americans are currently hospitalized with COVID-19. It seems like molnupiravir is arriving just when we need it.

It isn’t the only antiviral COVID pill, either. A day after the U.K. authorized Merck’s drug, Pfizer announced that its antiviral, Paxlovid, was also staggeringly effective at preventing the progression of COVID-19 in high-risk patients. The drug, when taken within three days of the onset of symptoms, reduced the risk of hospitalization by nearly ninety per cent. Only three of the nearly four hundred people who took Paxlovid were hospitalized, and no one died; in the placebo group, there were twenty-seven hospitalizations and seven deaths. Paxlovid is administered along with another antiviral medication called ritonavir, which slows the rate at which the former drug is broken down by the body. Like Merck, Pfizer is now examining whether Paxlovid can also be used to prevent infections after an exposure. Results are expected early in 2022. (It’s not yet known how much of a difference the drugs will make for vaccinated individuals suffering from breakthrough infections; Merck’s and Pfizer’s trials included only unvaccinated people with risk factors for severe disease, such as obesity, diabetes, or older age. Vaccinated individuals are already much less likely to be hospitalized or die of COVID-19.)

Paxlovid interrupts the virus’s replication not by messing with its genetic code but by disrupting the way its proteins are constructed. When a virus gets into our cells, its RNA is translated into proteins, which do the virus’s dirty work. But the proteins are first built as long strings called polypeptides; an enzyme called protease then slices them into the fragments from which proteins are assembled. If you can’t cut the plywood, you can’t build the table, and Paxlovid blunts the blade. Because they employ separate mechanisms to defeat the virus, Paxlovid and molnupiravir could, in theory, be taken together. Some viruses that lead to chronic infections, including H.I.V. and hepatitis C, are treated with drug cocktails to prevent them from evolving resistance against a single line of attack. This approach is less common with respiratory viruses, which don’t generally persist in the body for long periods. But combination antiviral therapy against the coronavirus could be a subject of study in the coming months, especially among immunocompromised patients, in whom the virus often lingers, allowing it the time and opportunity to generate mutations.

Merck will be producing a lot of molnupiravir. John McGrath, the company’s senior vice-president of manufacturing, told me that Merck began bolstering its manufacturing capacity long before the Phase III trial confirmed how well the drug worked. Normally, a company assesses demand for a product, then brings plants online slowly. For molnupiravir, Merck has already set up seventeen plants in eight countries across three continents. It now has the capacity to produce ten million courses of treatment by the end of this year, and at least another twenty million next year. It expects molnupiravir to generate five to seven billion dollars in revenue by the end of 2022.

How much will all these pills soften the looming winter surge? As has been true throughout the pandemic, the answer depends on many factors beyond their effectiveness. The F.D.A. could authorize molnupiravir within weeks, and Paxlovid soon afterward. But medications only work if they make their way into the body. Timing is critical. The drugs should be taken immediately after symptoms start—ideally, within three to five days. Whether people can benefit from them depends partly on the public-health infrastructure where they live. In Europe, rapid at-home COVID tests are widely available. Twenty months into the pandemic, this is not the case in much of the U.S., and many Americans also lack ready access to affordable testing labs that can process PCR results quickly.

Consider one likely scenario. On Monday, a man feels tired but thinks little of it. On Tuesday, he wakes up with a headache and, in the afternoon, develops a fever. He schedules a COVID test for the following morning. Two days later, he receives an e-mail informing him that he has tested positive. By now, it’s Friday afternoon. He calls his doctor’s office; someone picks up and asks the on-call physician to write a prescription. The man rushes to the pharmacy to get the drug within the five-day symptom-to-pill window. Envision how the week might have unfolded for someone who’s uninsured, elderly, isolated, homeless, or food insecure, or who doesn’t speak English. Taking full advantage of the new drugs will require vigilance, energy, and access.

Antivirals could be especially valuable in places like Africa, where only six per cent of the population is fully vaccinated. As they did with the vaccines, wealthy countries, including the U.S. and the U.K., have already locked in huge contracts for the pills; still, Merck has taken steps to expand access to the developing world. It recently granted royalty-free licenses to the Medicines Patent Pool, a U.N.-backed nonprofit, which will allow manufacturers to produce generic versions of the drug for more than a hundred low- and middle-income countries. (Pfizer has reached a similar agreement with the Patent Pool; the company also announced that it will forgo royalties for Paxlovid in low-income countries, both during and after the pandemic.) As a result, a full course of molnupiravir could cost as little as twenty dollars in developing countries, compared with around seven hundred in the U.S. “Our goal was to bring this product to high-, middle-, and low-income countries at fundamentally the same time,” Paul Schaper, Merck’s executive director of global pharmaceutical policy, told me. More than fifty companies around the world have already contacted the Patent Pool to obtain a sublicense to produce the drug, and the Gates Foundation has pledged a hundred and twenty million dollars to support generic-drug makers. Charles Gore, the Patent Pool’s executive director, recently said that, “for large parts of the world that have not got good vaccine coverage, this is really a godsend.” Of course, the same challenges of testing and distribution will apply everywhere.