Caribou Biosciences is shutting down. The company was co-founded by Jennifer Doudna, who shares a Nobel Prize for the gene editing chemistry the entire field now runs on, it raised about $350 million over its life, and it is closing because it could not fund the late-stage trial that would have told the world whether its CAR-T therapy works in lymphoma. Wind-down costs are put at $5 million to $8 million. The science is not being disproved. It is being shelved for want of the money to finish asking the question.
This is the specific and badly understood failure mode of therapeutic biotech. The curve of capital required does not flatten as a drug improves, it steepens. Discovery is cheap, relatively, and funded by people who enjoy discovery. Early clinical work is expensive and fundable on the strength of a signal. Late-stage registrational trials in cell therapy are a different order of magnitude, hundreds of patients across dozens of sites, manufacturing to commercial specification, years of follow-up, and a cash requirement that arrives precisely when the earlier investors have been diluted into exhaustion and the public markets have decided the sector is out of fashion. A company can do everything correctly and arrive at that wall with a promising asset and an empty account. Three hundred and fifty million dollars is a great deal of money and it is not enough to finish a phase three in oncology.
The coffee parallel is almost too neat, and roasters live it constantly. A grower can produce an extraordinary lot, genuinely competition grade, with the altitude, the varietal and the processing all exactly right, and still never get it in front of a buyer who would pay for it. What stands between the cherry and the cup is not agronomy. It is the mill, the certification, the export licence, the container, the sample shipped to the right importer in the right month. Every one of those is capital and logistics rather than quality, and the farms that fail are mostly not the farms that grow badly. They are the farms that grew something excellent and could not get it to the port.
Caribou could not get it to the port. What makes that genuinely costly is that the information the trial would have produced is now unowned. A late-stage dataset in allogeneic CAR-T, the off-the-shelf approach that would make cell therapy something other than a bespoke six-figure manufacturing exercise per patient, is a public good dressed up as a corporate asset. Whoever runs that trial next will spend the same years and the same money discovering what Caribou was positioned to discover in 2027, and the patients in that gap are the ones paying the interest.
The assets themselves will not evaporate. Intellectual property in gene editing has a strong resale market, programmes get picked up, and the people scatter into other companies carrying what they learned, which is how this sector has always recycled its failures. But a shutdown with $5 million to $8 million of wind-down cost against $350 million deployed is a reminder that the binding constraint in biotech is the financing calendar rather than the biology. Nobel-grade chemistry does not shorten a trial or soften a term sheet.
The scissors worked. They have worked since the day the prize was awarded. What ran out was the budget for the experiment that would have told us what they could cut.
Source: STAT News, “Caribou Biosciences Shutting Down, Abandoning Lymphoma CAR-T Therapy”
