Pharma
Deployment record
In the tracker →- Organisation
- Anthropic
- System
- Claude
- Stage
- Live
- Function
- drug development
- 01Mithril Medicine, a one-employee biotech company, was acquired by Anthropic in July, with its drug programs becoming the first in Anthropic's internal pipeline.
- 02Cody Tranbarger, the founder of Mithril, utilized Claude to analyze data and manage logistics across 11 outside labs for drug development.
- 03Mithril is focused on developing bispecific antibodies for LAMA2-related congenital muscular dystrophy and other genetic diseases, with potential human trials starting as early as late 2028.
Show all 5Show less
- 04Claude helped reduce human review time for manufacturing records from over 25 hours to less than three in collaboration with Enveda.
- 05OpenAI's acquisition of Orchestra Bio aims to enhance tools for drug developers, while Anthropic is leveraging Claude for its drug programs.
Anthropic
Follow Deployed on Google
When an outside lab sent results to Mithril Medicine, they landed with the same person who had to decide what to test next. Cody Tranbarger, the founder and only full-time employee of the Cambridge, Massachusetts, biotech, used Claude to analyze data and track experiments across 11 outside labs.
Anthropic acquired Mithril in July and disclosed the deal last month. In a LinkedIn announcement, Tranbarger said Mithril’s programs became the first in Anthropic’s internal pipeline and that he joined to lead the company’s broader drug development effort.
What Claude Took On
Outsourcing experiments predates generative AI. Biotech entrepreneur Matt Davidson told Nature that a company he started in 2013 operated without a lab of its own and had two full-time employees until its drug was ready for late-stage testing.
Mithril hired part-time consultants and paid at least five contract research organizations to run experiments. Tranbarger used Claude to help manage contracts and logistics, and called the setup “Pipeline-in-a-Person.”
The economics depend on whether one scientist can advance more programs or decide sooner which ones to stop. A lean payroll still leaves bills for consultants and lab work. Anthropic disclosed how long one program took, with no cost figures or timeline for a comparable program run without Claude.
The week's drug discovery, trials and health-system deployments. Wednesdays.
Anthropic CEO Dario Amodei pointed to the range of skills a founder needs in his recent interview with Walter Isaacson on The Builders. “One theory I have is that running a company is too big a task for any one person,” he said.
From Mouse Data to a Human Trial
Mithril set out to develop bispecific antibodies, which can bind to two targets, for people with LAMA2-related congenital muscular dystrophy (LAMA2-CMD) and other genetic diseases affecting the extracellular matrix. The mouse results came from the LAMA2-CMD program. If the candidates continue to hold up, Anthropic says a drug could be ready to start human trials as early as late 2028.
Getting to a trial would also mean manufacturing the drug for human testing and documenting how each batch was made. In a separate test with biotech company Enveda, Anthropic says Claude helped flag issues in manufacturing records and cut human review time from more than 25 hours to less than three.
Where OpenAI Fits In
A week before Anthropic disclosed the Mithril deal, OpenAI announced its acquisition of Orchestra Bio, a platform that helps drug developers track studies and coordinate work across teams. OpenAI is adding those tools to Rosalind for life sciences researchers, while Anthropic is using Claude in drug programs it now owns.
“I am extremely excited to see how far we can push AI to discover new science,” OpenAI CEO Sam Altman said in his latest interview with Raj Shamani on Figuring Out. He named finding new medical treatments as one of the goals.
For Tranbarger, the next test is whether the approach can improve the odds across several drug programs. “AI is poised to transform drug development; if we get it right, the result will be more programs, higher hit rates, and broader patient impact,” he said.
Last updated
