Recursion

Recursion is a drug discovery company that pairs automated laboratory experiments with computation. Rather than testing a small number of hand-picked hypotheses, it runs experiments at scale and uses the data they generate to train models that point toward possible medicines. The company describes its platform, Recursion OS, as spanning the whole path from target identification through to clinical trial enrolment ^[2].

Founding and early financing

Recursion was founded on 5 November 2013 ^[1]. It closed a Series A round in September 2016 led by Lux Capital, and a Series B in September 2017 led by DCVC ^[1]. Baillie Gifford led a Series C in February 2019 ^[1]. In September 2020 the company announced a partnership with Bayer alongside a $239 million Series D led by Leaps by Bayer ^[1].

Those rounds trace the usual path of a research-heavy startup: early venture money to build the technology, later rounds from larger investors to scale it. What is less usual is the pairing of financing with partnerships. In Recursion's case the same announcement carried both a corporate collaborator and an investor, so the capital and the scientific relationship arrived together ^[1].

Public listing and large-pharma collaborations

Recursion began trading on Nasdaq under the symbol RXRX after closing its initial public offering on 16 April 2021 ^[1]. Going public gives a drug discovery company a source of capital that does not depend on the timing of a single partnership or clinical result, at the cost of quarterly scrutiny.

In December 2021 the company announced a multi-year collaboration with Roche and Genentech in neuroscience and oncology ^[1]. Such agreements typically give a smaller platform company access to a large partner's clinical development machinery and disease expertise, while the partner gets access to the platform. Both sides are betting that the discovery method itself, rather than one compound, is the valuable asset.

The laboratory, the dataset and the computers

The company says its automated laboratories run up to 2 million experiments a week and have produced a proprietary dataset of more than 50 petabytes ^[2]. The argument for that scale is statistical. Machine learning models improve when they have many labelled examples, and in biology the supply of examples is limited by how fast experiments can be run and how consistently they are recorded. Automating the experiments addresses both problems at once: throughput goes up, and the resulting measurements are generated in a uniform way that makes them usable as training data.

Recursion has also bought in capabilities rather than building everything itself. In May 2023 it acquired Cyclica and Valence, adding digital chemistry and generative AI capabilities ^[1]. And in March 2024 it completed BioHive-2, which it describes as the largest supercomputer in the pharmaceutical industry ^[1]. The compute investment follows from the data strategy: models trained on petabyte-scale experimental data need machines to train them and to run the resulting predictions.

Read together, these moves describe a loop. Experiments produce data; data trains models; models propose the next experiments; automation runs them. The company's own description of Recursion OS as covering everything from target identification to clinical trial enrolment is a claim about how far around that loop it intends to operate ^[2].

Merger with Exscientia and a change of leadership

On 20 November 2024 Recursion completed its combination with Exscientia ^[1]. Combining two organisations that both work at the intersection of computation and drug discovery means merging not only pipelines but also software, laboratory operations and scientific staff — the part of such deals that takes longest to settle.

On 1 January 2026 Najat Khan became chief executive and president, with co-founder Chris Gibson moving to chairman of the board ^[1]. Having a founder step from the executive role to the board while an outsider takes over operations is a common transition for a company moving from building a platform to running a larger, merged organisation.

Where it sits

Recursion's history is a compact record of one bet in drug discovery: that biology can be treated as a data problem, and that the company which generates the most and the cleanest experimental data will find medicines faster than one which relies on individual insight alone. The funding rounds, the listing, the academic-style collaborations with large pharmaceutical companies, the acquisitions and finally the merger are all steps toward making that bet at a larger scale. Whether the scale translates into approved medicines is a question the company's own materials do not answer; what they document is the build-out — laboratories, datasets, computers and partnerships — that the answer will depend on.