Recursion

When Biology Becomes a Computation Problem

Drug discovery has always been constrained by scale. Biology is staggeringly complex, experiments are slow and expensive, and success rates are painfully low. For decades, progress depended on narrowing questions—testing one hypothesis at a time, advancing a small number of compounds through years of trial and error. Recursion was founded on the belief that this approach no longer matches the tools available to science.

Instead of treating biology as a sequence of isolated experiments, Recursion treats it as a data-rich system—one that can be mapped, modeled, and interrogated computationally. The company’s ambition is not to accelerate individual experiments, but to industrialize the discovery process itself.

From Wet Lab to Data Engine

Recursion’s core insight is that biology can be observed at scale if experiments are designed to generate data systematically. The company operates automated wet labs that run millions of experiments, capturing high-dimensional biological data through imaging and molecular readouts.

Cells are perturbed—genetically or chemically—and their responses are recorded using advanced microscopy. Each experiment generates rich visual data that reflects how cells change under different conditions. Over time, this creates a massive, structured dataset linking genes, compounds, and phenotypes.

Recursion has described this as building a “biological map”—a representation of how perturbations propagate through cellular systems. That map becomes the substrate for machine learning models trained to identify patterns humans would struggle to see.

Machine Learning Meets Experimental Biology

What distinguishes Recursion from many AI-in-biotech efforts is the tight integration between experimentation and computation. Data is not an output of the process; it is the fuel that drives it.

The company uses machine learning to analyze cellular images, embedding biological responses into numerical representations that can be compared, clustered, and searched. This allows Recursion to identify relationships between diseases, genes, and potential treatments in a systematic way.

Chris Gibson, Recursion’s co-founder and CEO, has framed the company’s approach succinctly: “We’re trying to turn biology into something that can be navigated like a data science problem.” That framing captures the company’s ambition—to move from intuition-driven discovery to computation-driven exploration.

Importantly, Recursion does not claim that AI replaces biology. Instead, it augments it. Hypotheses are still tested experimentally, but they are generated and prioritized by models trained on unprecedented volumes of biological data.

Scale as a Strategic Advantage

Recursion’s emphasis on scale is deliberate. Many biotech companies apply machine learning to small datasets or narrow targets. Recursion’s platform is designed to grow continuously, with each experiment expanding the dataset and improving model performance.

This creates a compounding effect. As more data is generated, predictions become more accurate. As predictions improve, experiments become more targeted. Over time, the platform becomes more valuable—not just for a single program, but across the entire pipeline.

The company has reinforced this strategy through partnerships and acquisitions aimed at expanding both data and compute capacity. In public announcements, Recursion has emphasized that its long-term advantage lies in owning and generating proprietary biological data rather than relying solely on public datasets.

From Platform to Pipeline

While Recursion often emphasizes its platform, it is also a drug development company with an active pipeline. The company advances programs in areas such as rare diseases, oncology, and neuroscience—fields where traditional discovery approaches face significant challenges.

What’s notable is how Recursion uses its platform to explore these areas. Instead of starting with a single target, the company often identifies multiple candidate compounds linked to disease-relevant phenotypes. This increases optionality and reduces dependence on any one hypothesis.

Recursion has also pursued partnerships with large pharmaceutical companies, leveraging its platform to support external discovery efforts while retaining control over its core technology. These collaborations serve both as validation and as a way to expand the platform’s reach.

Compute as a First-Class Citizen

Recursion’s approach requires immense computational resources. Training models on millions of high-resolution biological images is not trivial. Recognizing this, the company has invested heavily in compute infrastructure.

In 2023, Recursion announced a partnership with NVIDIA to build one of the largest supercomputers dedicated to biological research. The goal was to accelerate model training and enable more complex simulations of biological systems. This move signaled that Recursion views compute not as an operational expense, but as strategic infrastructure.

By aligning advances in biology with advances in compute, Recursion positions itself at the intersection of two rapidly evolving domains.

A Different Biotech Culture

Culturally, Recursion looks different from traditional biotech firms. Its teams include biologists, data scientists, engineers, and roboticists working side by side. Experiments are automated, data flows continuously, and iteration cycles are measured in days rather than months.

This culture reflects a broader shift in life sciences toward convergence with software engineering. Version control, continuous integration, and systematic testing—concepts familiar to technologists—are increasingly relevant to experimental science at scale.

Recursion has argued that this convergence is not optional. As biological datasets grow larger and more complex, manual approaches simply cannot keep up.

Risks and Realities

Recursion’s ambition does not eliminate risk. Drug discovery remains uncertain, and no platform can guarantee success. Translating computational insights into safe, effective medicines is a long and regulated process.

The company has been careful not to oversell AI as a shortcut. In public communications, Recursion emphasizes that its goal is to improve probabilities, not eliminate failure. This realism is important in an industry where exaggerated claims can erode trust.

Why Recursion Belongs in Rewired 100

Recursion represents a fundamental shift in how one of humanity’s most complex challenges is approached. By treating biology as a data system, it is attempting to rewire drug discovery from the ground up.

In the context of Rewired 100, Recursion stands out because it is not building an application—it is building infrastructure for science. Its platform does not promise instant cures; it promises a better way to explore biology, systematically and at scale.

As AI continues to move into the physical and biological world, the companies that matter most will be those that respect complexity while harnessing computation. Recursion’s work suggests that the future of medicine may be shaped less by isolated breakthroughs and more by platforms that make discovery itself more navigable.

If that future arrives, it will not look like a single moment of insight. It will look like millions of experiments, connected by models, quietly revealing patterns that were always there—waiting to be computed.