Anthropic unveiled Claude Science on June 30, 2026, at a San Francisco event, marking the company's decision to elevate the product to the same rank as Claude Code and Claude Cowork, a signal of how seriously it takes AI's role in scientific research. Alongside the product launch, Anthropic announced it will use Claude Science to pursue its own research into drugs for rare, neglected diseases, stepping directly into drug development rather than only selling tools to others. Coverage from outlets including
CNBC,
MIT Technology Review, and
ZME Science frames this as Anthropic's most significant move yet into life sciences.
On the technical side, Claude Science ships as a standalone product with more than 60 curated skills and connectors pre-configured for genomics, protein biology, and chemistry, able to interface with the computational tools molecular biologists and drug hunters use daily, run code on high-performance computing clusters, and trace results back to their source for reproducibility. This builds on earlier groundwork: in October, Anthropic had already released plug-ins under the heading "Claude for Life Sciences" to help Claude use scientific software and databases, but Claude Science is a fuller, standalone product. During the launch, Alexander Tarashansky, who led the product's development, demonstrated how Claude Science could autonomously identify new drug candidates for phenylketonuria, a rare genetic disease caused by a deficiency in an enzyme that metabolizes phenylalanine — a concrete proof-of-concept rather than a purely conceptual pitch.
Why start with rare and "neglected" diseases specifically? Many rare disorders stem from a single damaged gene, which can make the biological cause easier to pin down than in common illnesses like Alzheimer's, diabetes, or heart disease, where many genes, tissues, and environmental factors interact. Anthropic's own leadership has been candid about the commercial logic behind this choice. Eric Kauderer-Abrams, Anthropic's head of life sciences, said the company will focus on discovering treatments for diseases that traditional biopharmaceutical companies wouldn't consider attractive targets, while Jonah Cool, Anthropic's head of life sciences partnerships, framed the effort as running alongside the AI tools Anthropic is selling to the drugmakers it hopes to court as customers.
The internal program is explicitly framed as both humanitarian and strategic. Anthropic describes it as an internal preclinical drug discovery effort aimed at neglected and rare diseases, meant to gather real lab experience that feeds back into the Claude Science workbench, in the same way Claude Code learned from real coding work. As a public benefit corporation, Anthropic can choose programs based on patient benefit rather than commercial return, and neglected diseases represent a genuine gap in the pharmaceutical pipeline — though the arrangement also gives Anthropic a live testbed to refine a commercial product it plans to sell to the pharmaceutical industry broadly.
Notably, the move follows Anthropic's April 2026 acquisition of computational biology firm Coefficient Bio for roughly $400 million, and the addition of Novartis CEO Vas Narasimhan to its board — details reported by Digg and echoed across the tech press, and which some commentary (including the Activist Post/Natural News piece you linked) treats as evidence of deepening ties to the pharmaceutical industry rather than a clean break from it. It's worth noting that piece is an opinion article, not a news report, and its framing — that "superintelligence" could just as easily be used to manufacture new lifelong customer bases as to cure disease — is speculative commentary rather than something reported by Anthropic or the primary outlets covering the launch. The underlying facts (the June 30 launch, the phenylketonuria demo, the Coefficient Bio deal, the neglected-disease focus) are well corroborated across CNBC, MIT Technology Review, and ZME Science; the darker "Big Pharma trojan horse" narrative is one interpretive lens among several, and it's reasonable to hold it more loosely than the reported facts themselves.
