What Anthropic’s S-1 Reveals About Its Plans For Scientific Research
By Alex Knapp,
Senior Editor.
On the first page of AI company Anthropic’s IPO filing, submitted to the U.S. Securities and Exchange Commission (SEC) today, the company highlights its goal of using AI to help with scientific research. The introduction says, “We have started to directly make biological discoveries with Claude.” It also says Anthropic expects Claude to make more scientific breakthroughs in the future.
The company is referring to its own biology research, which it announced last month. In that work, its software discovered a new family of enzymes with features similar to those of enzymes used in CRISPR gene editing. Anthropic has also built several products for scientists. Claude Science lets scientists study data and create figures or charts that are ready to use in published research. Claude for Life Sciences is designed to help researchers discover drugs and prepare documents for submissions to the U.S. Food and Drug Administration (FDA), along with similar paperwork. The filing also says Anthropic has developed software that lets AI agents control lab equipment, including microscopes, machines that handle liquids, and standard robotic arms. These tools can help automate scientific research.
Anthropic reported $16.3 billion in revenue and an operating loss of $693 million during the first half of 2026. But its filing does not say how much revenue came specifically from its research products. It names customers such as pharmaceutical company Bristol Myers Squibb, the Garvan Institute of Medical Research, and the University of California, San Francisco. Anthropic has also attracted attention through partnerships with companies including Lilly and Novo Nordisk. Earlier this year, NASA used Claude Code to help guide the Mars rover Perseverance during a mission.
Not all of Anthropic’s life sciences tools are meant for long-term research. The company says the World Health Organization has used Claude during the ongoing Ebola outbreak to reduce the time needed to prepare its daily situation report from a full day to less than an hour. The organization has also used the AI to run several disease models and help decide where to place treatment centers. Anthropic says it is also working with nonprofit groups, including the Gates Foundation and the International Rescue Committee, to study health data. This work aims to help detect disease outbreaks and find the most effective ways to send healthcare workers where they are needed.
Using Anthropic’s AI models for research could bring major benefits, but it also carries serious risks. One concern is that the models can operate robots and laboratory equipment. Anthropic warns that mistakes caused by a model's behavior or deliberate misuse could “result in bodily injury, property damage, damage to costly equipment or experimental materials, or the release of hazardous substances.” The company also warns that its models could help people who want to develop chemical, biological, or other dangerous weapons. Anthropic co-founder and CEO Dario Amodei has warned about this danger for years.
Anthropic says it has put technical protections in place to stop people from using its tools to create weapons. However, it warns that these protections are not complete. They may become outdated as threats change, people may find ways around them, and they may not always work as intended.