Pfizer is a household name thanks to its COVID-19 vaccine, of which it expects to produce as many as 3 billion doses this year. That would translate into $26 billion in sales, which would make it the "biggest-selling medicine ever," reports Christopher Rowland in a lengthy piece for the Washington Post that details how getting to this point was far from a sure thing. In fact, its starting point was "utter failure." But before they got to that failed trial run at a Pfizer plant outside Kalamazoo, Mich., came the March 20, 2020, decision to go with that Kalamazoo plant as the place that would "formulate doses, fill vials and ship boxes of vaccine packed with dry ice," explains Rowland, who writes that happened in a quarter-mile long building that didn't initially contain the sizable machines needed to combine lipid nanoparticles and then filter the result—because they didn't exist.
By Sept. 11, plant operators were ready to do a trial run, but something very quickly went wrong. "When operators checked a vat at the end of the production run, most of the key ingredient—the fat molecule encasing the messenger RNA—was missing," writes Rowland. They identified a membrane as the point of failure and established an integrity test that would ensure the membranes weren't allowing lipid nanoparticles to leak. It worked, but the fine-tuning continued (Rowland notes, for instance, that they created a system that would allow it to reuse filters to get around shortages). On Dec. 13, the first vials started shipping—and on that same day, the plant's head of operations was told he had to figure out how to up production 4x by the following week. (Read the full story for more on how they tackled that complex challenge.)