Anthropic partners with biotech firms to use AI for drug discovery

Accelerating the lab workflow

Anthropic is expanding its role in the medical field by partnering with six different organizations to explore how artificial intelligence can speed up drug discovery. The company, which created the AI chatbot Claude, is working with entities like Bristol Myers Squibb and BigHat Biosciences. These collaborations aim to use AI to identify potential drugs and design molecules more efficiently than traditional methods.

The primary goal of these partnerships is to reduce the significant time and cost barriers that currently exist in medical research. By using AI to navigate complex biological data, researchers hope to move from the initial discovery phase to actual treatments much faster. This approach targets a variety of conditions, including cancer, food allergies, and rare diseases.

While the company has made bold statements about the future of medicine, it remains grounded in the current reality of scientific testing. Anthropic notes that much of this research is still in the preclinical stage. This means the treatments have not yet been tested in humans and must still pass rigorous laboratory benchmarks before they can be used in patients.

Redesigning cancer therapies

One specific project involves working with Generate Biomedicines to adapt CAR T-cell therapy. This type of treatment is traditionally used to fight blood cancers, but the goal here is to make it effective against solid tumors. Scientists used Claude Code to propose a redesigned version of a section of the therapy's genetic instructions.

The AI-generated design was approximately 70 per cent shorter than the standard version used in previous therapies. When researchers tested this new design in a laboratory, the results were significant. The shorter instructions helped immune cells produce roughly 10 times more receptors compared to the original version.

In addition to increasing receptor production, the redesign improved the ability of the immune cells to kill tumor cells. The tests also showed that these cells were able to remain active for a longer period of time. These results suggest that AI can help refine complex biological instructions to create more potent and durable medical interventions.

Rapid antibody and molecule design

Anthropic is also using its technology to tackle food allergies through a collaboration with Divergence Labs. In just about one month, the partnership generated around 35,000 antibody designs. This scale of production highlights how AI can handle massive amounts of variation that would take humans much longer to process manually.

Working with BigHat Biosciences, the team looked at enhancing two potential drugs, including one meant for cancer cells. After evaluating roughly 1,010 different versions, the researchers identified a top candidate that attached to its intended target with a strength exceeding 10 times that of the original.

Beyond the design of the molecules themselves, Anthropic is helping to streamline the administrative side of drug manufacturing. In a project with Enveda, the company assisted in cutting the duration of manual checks for production paperwork by 89 per cent. This shows that AI can optimize both the high-level science and the logistical paperwork of the pharmaceutical industry.

Targeting rare childhood diseases

In addition to its corporate partnerships, Anthropic has launched its own internal drug discovery programs. One of these initiatives focuses on rare childhood forms of muscular dystrophy. By developing its own pipeline, the company is moving from a service provider to a direct participant in medical research.

The company stated that a specific candidate for muscular dystrophy might be prepared for human clinical testing by the end of 2028. However, this timeline is dependent on the promising results from current tests continuing to hold up. If the data remains consistent, this would mark a major milestone for the company's independent research efforts.

The move into rare diseases highlights a specific niche where the high costs of development often prevent traditional pharmaceutical companies from investing. By using AI to lower the costs of research, Anthropic aims to make it more feasible to develop treatments for conditions that affect smaller populations of people.

The path to human trials

The transition from AI-generated designs to actual medicine requires a bridge of physical testing. While the AI can propose thousands of designs in a month, each one must be synthesized and tested in a lab environment. These preclinical tests are essential to ensure that the molecules are safe and effective before they ever reach a person.

The collaboration with various biotech firms provides a structured way to move these designs through the regulatory pipeline. By working with established organizations like Bristol Myers Squibb, Anthropic can leverage existing infrastructure for clinical trials. This partnership model allows the AI to act as a powerful engine for discovery while the partners provide the necessary oversight and testing.

As these projects move toward 2028, the focus will shift from generating designs to proving their efficacy in living systems. The success of these initiatives will depend on whether the AI-optimized molecules can perform as well as, or better than, those developed through traditional methods. For now, the technology serves as a massive accelerator for the scientists on the front lines of medicine.