What's Happening

Anthropic, the artificial-intelligence company behind Claude, has quietly established a physical biology laboratory in the San Francisco Bay Area as it expands its work in life sciences and drug research.

The company has been increasingly moving beyond computer-based, or "in silico," biological research toward experiments involving physical samples and laboratory equipment. Anthropic's head of life sciences, Eric Kauderer-Abrams, confirmed to Reuters that the company has a wet lab and is already conducting physical biology work there. (MarketScreener)

A company spokesperson later clarified that the facility is not specifically for drug discovery, although Anthropic has said it wants to develop tools and eventually run preclinical programs aimed at diseases that traditional pharmaceutical companies may not find financially attractive. (MarketScreener)

The move marks a significant expansion of Anthropic's role in healthcare and biotechnology. Instead of providing AI tools only to scientists and pharmaceutical companies, Anthropic is beginning to build some of the physical infrastructure needed to conduct biological research itself.

Anthropic Is Moving From AI Software Toward Physical Biology

Much of AI-driven drug discovery has historically taken place on computers.

Researchers can use AI to analyze datasets, predict protein structures, identify potential drug targets, design molecules and prioritize experiments. But those predictions ultimately have to be tested in the real world.

Anthropic says it views physical experimentation as an essential part of biology.

Kauderer-Abrams told Reuters that the final test in biology remains real laboratory work and said Anthropic is already carrying out some of that work internally while using external partners for other experiments. (MarketScreener)

The company therefore appears to be developing a hybrid model:

  • AI models generate and analyze biological hypotheses.
  • Software and automation help translate those ideas into experiments.
  • Physical laboratories test the resulting biological predictions.
  • Human scientists and external partners continue to provide oversight and specialized capabilities.

That approach moves Anthropic closer to the way a biotechnology organization operates, even though the company says it is not currently trying to become a conventional pharmaceutical developer.

The Lab Is Part of a Much Larger Life-Sciences Strategy

The new laboratory is not an isolated investment.

Anthropic has spent much of 2026 building a broader AI-for-science and life-sciences business.

In June, the company launched Claude Science, a scientific AI workbench designed to connect researchers with scientific databases, computing tools and specialized workflows. Anthropic says the platform supports areas including genomics, single-cell biology, proteomics and cheminformatics. (Anthropic)

In August, Anthropic also introduced the Model Hardware Standard, a system intended to allow AI agents to interact with laboratory and manufacturing equipment such as microscopes, liquid handlers and robotic arms through a common interface. (Anthropic)

And on September 17, Anthropic launched its Life Sciences Verification Program, giving vetted life-sciences organizations access to its models with safeguards designed for biology-related work including drug discovery, research biology, clinical development and manufacturing. (Anthropic)

Taken together, those developments show that Anthropic is building an ecosystem around AI-assisted biological research rather than treating healthcare as a small standalone application.

Anthropic Says Rare Diseases Are a Major Focus

One of the company's stated ambitions is to work on rare diseases and other conditions that may not receive enough investment from traditional drug developers.

Kauderer-Abrams said Anthropic plans to conduct preclinical work in areas where conventional companies may not see a sufficiently attractive financial return. (MarketScreener)

This reflects a potential economic use case for AI in pharmaceutical development.

Drug development can be extremely expensive, and some diseases affect relatively small patient populations. A company may therefore face difficulty justifying the cost of extensive research, discovery and clinical development for certain rare conditions.

Anthropic believes AI could lower some of the costs and time involved in those early stages.

However, the company has not publicly disclosed the specific diseases it is targeting or demonstrated that it has produced a drug candidate ready for human testing. Reuters noted that the exact diseases and the company's progress remain unclear. (MarketScreener)

The Goal Includes Previously Difficult-to-Treat Diseases

Anthropic sees AI as potentially useful for conditions that have historically been difficult to address.

Kauderer-Abrams pointed to complex biological molecules, including bispecific and trispecific antibodies, as examples of areas where AI could potentially accelerate development.

These molecules can be designed to interact with multiple biological targets, making them useful in some complex diseases. (MarketScreener)

The underlying idea is that more capable AI systems could improve several steps in the discovery process, including:

  • Finding biological targets: AI can process large quantities of biological information and help researchers identify potential disease mechanisms or targets.
  • Designing molecules: Models can generate and evaluate potential protein or molecular designs much faster than conventional trial-and-error processes.
  • Planning experiments: AI can help researchers determine which experiments are most informative and which variables should be tested next.
  • Interpreting results: AI can analyze large experimental datasets and help researchers identify patterns that may otherwise take substantial time to uncover.

Anthropic has recently published its own research showing Claude being used for protein design and biomolecular modeling. The company said Claude designed protein binders against 15 targets and succeeded against 14, with individual experimental success rates ranging from 22% to 35% depending on the setup. (Anthropic)

Those are research results, not evidence of a successful human drug-development program.

Anthropic Is Building Toward Automated Laboratories

A particularly important part of the company's strategy is laboratory automation.

Anthropic wants Claude to do more than analyze experimental data. It is working toward systems in which AI agents can interact directly with robotic laboratory equipment and coordinate experiments.

The company says its Model Hardware Standard can allow AI agents to operate multiple instruments, including liquid handlers, microscopes and robotic arms, through a standardized interface. (Anthropic)

In early work described by Anthropic, AI-controlled systems were able to coordinate laboratory equipment and execute experiments with less manual programming.

Anthropic's stated goal is not to remove scientists from the laboratory altogether. Instead, it says AI should handle repetitive and mechanical parts of experimental execution while scientists focus on experimental design, interpretation and decisions requiring human judgment. (Anthropic)

Human Oversight Remains Important

The expansion into physical biology comes with a major challenge: AI systems can make errors in the physical world.

Anthropic says human oversight remains essential when AI interacts with laboratory equipment.

Its own research on the Model Hardware Standard acknowledges limitations in Claude's understanding of the physical world. The company says researchers have had to guide Claude in recognizing physical laboratory failures, such as problems caused by foaming in biological samples, that an AI might otherwise interpret as software issues. (Anthropic)

Anthropic therefore sees laboratory automation as an early-stage effort rather than a fully autonomous scientific system.

Kauderer-Abrams told Reuters that the company is still in the early stages of using AI to automate laboratory work. (MarketScreener)

The Timing Is Significant

Anthropic's expansion into biology is happening at a complicated moment for the AI industry.

The company has simultaneously been warning about the potential risks of increasingly capable AI systems.

Reuters reported that Anthropic researchers had recently raised concerns about advanced AI potentially contributing to catastrophic outcomes, while the company had also identified scenarios where its systems could potentially be misused for biological-weapons development. (MarketScreener)

That creates a tension for the company.

The same technologies that may help scientists accelerate drug research could also create new risks when applied to biological systems.

Anthropic says its life-sciences strategy therefore requires balancing greater biological capability with stronger safeguards and human oversight. (MarketScreener)

Anthropic Is Already Working With Major Pharmaceutical Companies

Anthropic is not entering healthcare from scratch.

Its AI services are already being used by major pharmaceutical companies, including Roche's Genentech, Bristol Myers Squibb and Novo Nordisk, according to Reuters. (MarketScreener)

That creates an unusual position for the company.

Anthropic can simultaneously act as:

  • A technology supplier to pharmaceutical companies.
  • A research platform for scientists.
  • A developer of AI-enabled laboratory infrastructure.
  • A potential participant in biological research itself.

This combination creates opportunities but also creates questions about competition and trust.

Pharmaceutical Customers Could Have Concerns About Competition

One issue identified by Reuters is whether pharmaceutical customers will be comfortable with Anthropic conducting its own biological research while also providing AI systems to competing drug companies.

A pharmaceutical company may use Anthropic's tools to analyze proprietary research, develop drug candidates or design experiments.

Even if customer data is technically separated, companies could still worry about whether an AI provider conducting its own biology research could gain knowledge about the broader industry or become a future competitor. (MarketScreener)

Anthropic says it walls off customer data and has created a boundary around its own biological activities.

Kauderer-Abrams told Reuters that the company is not competing with pharmaceutical and biotechnology companies whose business is bringing drugs to market. (MarketScreener)

For now, Anthropic says it is not running clinical trials and is focusing on areas that industry may not otherwise address. (MarketScreener)

Anthropic Has Acquired Biological Expertise

The company has also been acquiring capabilities rather than building everything internally.

Anthropic acquired Coefficient Bio, a startup focused on AI and biology, in a stock transaction that was reported at roughly $400 million. Anthropic confirmed the acquisition and said it would help build tools for drug development, but it did not comment on the reported transaction value. (MarketScreener)

Anthropic has also added Novartis CEO Vas Narasimhan to its board, giving the company another senior connection to the global pharmaceutical industry. (MarketScreener)

These moves suggest the company is attempting to combine AI expertise with deeper knowledge of pharmaceutical research and biological development.

The Company Is Hiring for Laboratory Operations

Anthropic's investment is also becoming visible through hiring.

The company has sought people who can lead procurement and laboratory operations, while another role focused on expertise in protein and nucleic acid characterization.

Anthropic has also said its goal is to accelerate progress in life sciences by an order of magnitude. (MarketScreener)

Building a physical lab requires very different capabilities from operating a software company.

The company needs laboratory equipment, supply chains, biological materials, specialized scientific staff, safety systems and procedures for managing experiments.

The hiring activity indicates that Anthropic is building those capabilities rather than simply outsourcing all biological work.

There Is Still a Long Way From Lab Experiment to Approved Drug

The existence of a wet lab should not be confused with having an AI-developed drug close to market.

A successful biological experiment is only an early step.

A drug candidate typically needs:

  • Preclinical research: Researchers must establish biological activity and gather safety information.
  • Clinical development: The drug then needs to be tested in humans through multiple stages of clinical trials.
  • Regulatory review: Evidence must be submitted to regulators such as the FDA.
  • Manufacturing and commercialization: Even an approved drug must be manufactured at scale and brought into healthcare delivery systems.

Reuters noted that Anthropic has not yet taken on the clinical-trial challenge. (MarketScreener)

This means any claims about AI dramatically shortening the entire drug-development timeline remain future possibilities rather than established outcomes.

Anthropic's AI Research Is Advancing Alongside the Lab

The physical lab also comes as Anthropic continues improving Claude's scientific capabilities.

On September 17, the company said Claude had optimized more than 30 open-source biomolecular models in less than four weeks, improving their speed by roughly four times on average. It also announced a protein-design competition with wet-lab validation for more than 5,000 designs. (Anthropic)

Those developments show how Anthropic is trying to connect computational improvements with real-world biological testing.

The company is effectively building both sides of the system: better models and a way to test what those models produce.

What This Means for Pharmaceutical R&D

The most important potential impact is on the economics and speed of early-stage research.

If AI can reduce the time needed to design molecules, select experiments, analyze results and operate laboratory equipment, pharmaceutical companies could potentially test more ideas with the same resources.

That could be particularly relevant to:

  • Rare diseases
  • Complex proteins
  • Difficult biological targets
  • Early-stage discovery
  • Experimental optimization
  • High-throughput screening

The impact is still uncertain because improvements in computational or laboratory speed do not necessarily translate into successful medicines.

Drug development remains constrained by biological complexity, clinical safety, patient recruitment and regulatory requirements.

What This Means for Anthropic's Business Model

Anthropic's life-sciences expansion may also diversify how it makes money.

Instead of selling only general-purpose AI models, the company is developing specialized products, scientific tooling, laboratory infrastructure and potentially research capabilities.

Claude Science is already positioned as a dedicated scientific workbench, while the Life Sciences Verification Program is designed specifically for organizations conducting biological and pharmaceutical work. (Anthropic)

The wet lab represents another step toward deeper vertical integration.

Rather than simply selling software to scientists, Anthropic is increasingly building technology that connects AI models, scientific data, laboratory hardware and real-world experiments.

Why This Matters

Anthropic's new biology lab is significant because it shows how AI companies are moving deeper into the physical process of scientific discovery.

Until recently, much of the AI-for-drug-development story centered on software: analyzing literature, predicting structures, searching databases and generating molecular designs.

Anthropic is now adding another layer: actually conducting physical biological experiments and developing systems that can connect AI agents with laboratory equipment. (MarketScreener)

At the same time, the company is not yet a conventional pharmaceutical developer. It has not publicly identified a specific drug candidate entering clinical testing, and Reuters reported that its precise disease programs and progress remain unclear. (MarketScreener)

The significance therefore lies less in an immediate new medicine and more in a possible change to how drug research is performed.

Looking Ahead

Anthropic's next steps are likely to involve expanding its internal biological capabilities, increasing laboratory automation and using AI models to support more sophisticated experimental workflows.

The company says life sciences is already one of its largest areas of investment by headcount and resources. (MarketScreener)

The key questions will be whether AI can translate faster experimentation into genuinely better drug candidates, whether automated laboratory systems can operate reliably at scale and how Anthropic manages the safety and competitive concerns that come with working directly in biology.

The company's boundary around clinical development is also important. For now, Anthropic says it is not running clinical trials and intends to work on areas the pharmaceutical industry may not address. (MarketScreener)

What This Means for Healthcare Marketers

For healthcare marketers, this development shows that the AI market in healthcare is expanding beyond software and analytics.

The opportunity now extends into laboratory automation, research infrastructure, biomolecular modeling, diagnostics and drug development.

For B2B healthcare companies, signals such as AI partnerships, scientific platform launches, lab automation investments, biotech acquisitions and research hiring can reveal where pharmaceutical organizations are increasing spending before those investments appear in conventional market data.

Anthropic's expansion also suggests that the buyers for healthcare AI may increasingly include not only IT and digital-health teams, but also R&D, laboratory, clinical-development and scientific leadership teams.

Key Takeaways

  • Anthropic has established a wet biology lab in the San Francisco Bay Area and is conducting physical biological experiments. (MarketScreener)
  • The company says the lab is not specifically for drug discovery, although it is expanding its broader drug-science and life-sciences program. (MarketScreener)
  • Anthropic says it wants to pursue areas such as rare diseases and difficult-to-treat conditions that may receive less investment from traditional drug developers. (MarketScreener)
  • The company has launched Claude Science, a scientific AI workbench, and a Life Sciences Verification Program for vetted biological research organizations. (Anthropic)
  • Anthropic is also developing the Model Hardware Standard, which is designed to let AI agents interact with laboratory equipment such as robotic arms, liquid handlers and microscopes. (Anthropic)
  • Anthropic acquired Coefficient Bio, reportedly for about $400 million in stock, to strengthen its drug-development capabilities. (MarketScreener)
  • The company already provides AI services to pharmaceutical companies including Genentech, Bristol Myers Squibb and Novo Nordisk. (MarketScreener)
  • Anthropic says it is not currently running clinical trials, meaning its work remains primarily in research and preclinical stages. (MarketScreener)
  • The company says human oversight remains important because AI systems still have limitations when operating in physical laboratory environments. (Anthropic)
  • The long-term commercial impact will depend on whether AI-driven biological research can produce viable drug candidates and ultimately translate into safer, faster and more economical drug development.