Recursion Pharmaceuticals is a biotechnology company using artificial intelligence and massive computing power to automate the discovery of new medicines. It generated $70 million in revenue during 2025 and currently has five drug candidates in human clinical trials. While most drug companies rely on human intuition to pick drug targets, Recursion uses a proprietary supercomputer called BioHive-2 to run millions of automated experiments every week, searching for patterns that the human eye would miss.
The investment thesis on Recursion is that its "Recursion OS" platform can turn the high-risk gamble of drug discovery into a predictable, repeatable engineering process. By building a digital map of human biology that expands with every experiment, the company aims to find better drugs faster and with a lower failure rate than traditional biotech firms.
We think Recursion has built the most credible technical foundation in the industry, but the business remains a high-stakes "show-me" story until an AI-discovered drug reaches the final stages of approval. The company is moving faster than its peers, but it is still years away from selling a finished product. The thesis breaks if the current clinical trials fail to show a clear benefit, as that would call into question whether the AI is actually finding better medicines or just finding them faster.
Recursion Pharmaceuticals stock crashed after its initial excitement faded and has stayed stuck in a long slump. It is down about 90% from five years ago as the company continues to burn through cash while trying to prove that its supercomputers can successfully invent new medicines. Lately, the price has perked up slightly as investors track its progress.
What does it do?
Recursion Pharmaceuticals is a growth-stage biotechnology business that earns money by discovering new drugs using a combination of laboratory automation and artificial intelligence. Instead of researching one disease at a time, the company uses its "Recursion OS" to run millions of experiments across thousands of human cell types. It generates revenue through two paths: a "wholly owned" pipeline where it keeps 100% of the profits from drugs it develops, and "partnerships" where large drug companies like Roche or Bayer pay Recursion for access to its data and supercomputing tools.
Where does revenue come from?
Most current revenue comes from milestone payments and research fees paid by large pharmaceutical partners. These partnerships with companies like Roche, Genentech, and Sanofi involve upfront payments followed by "milestones" when specific research goals are met. In 2025, the company brought in $70 million, primarily from these collaborative agreements.
Who are its customers?
Recursion primarily serves giant pharmaceutical corporations and institutional research partners while advancing its own internal drug pipeline. Its partner list includes industry leaders like Roche, Genentech, Bayer, and Sanofi, who collectively provide the majority of the company's current cash inflows. As of the first quarter of 2026, Recursion reported $6.5 million in revenue, which was lower than the $14.7 million in the prior year due to the timing of completed research phases for Roche. The company also collaborates closely with NVIDIA, which provided $50 million in investment and technical support to build the BioHive-2 supercomputer, the most powerful AI system wholly owned by a pharmaceutical company.
What gives it staying power?
Recursion’s staying power comes from its proprietary dataset of over 25 petabytes of biological images and its massive NVIDIA-powered supercomputing infrastructure. This "BioHive-2" system allows the company to run and analyze experiments four times faster than its previous systems. Because the company owns the data from every experiment it has ever run, it has built a "map of biology" that competitors cannot easily replicate without spending years in the lab.
Where is it headed?
The company is focused on moving its lead drugs into "registrational" trials, which are the final clinical steps before seeking government approval for sale. Management is currently engaging with the FDA to define the path forward for REC-4881, a drug for a rare genetic condition that recently showed a 43% reduction in polyp burden in patients. If these internal drugs succeed, Recursion will transition from a research platform into a commercial drug seller.
Revenue is currently volatile and dependent on the timing of research milestones rather than steady sales. The $70 million generated in 2025 was a 17% increase over the prior year, but Q1 2026 revenue dropped to $6.5 million as large project phases with Roche were completed. This pattern is normal for a platform-based biotech, where cash arrives in large, lumpy payments when specific scientific goals are reached.
Free cash flow is negative as the company reinvests heavily into its supercomputing core and clinical trials. The company burned $81.1 million in cash during the first quarter of 2026, an improvement from the $132 million burned in the same period last year. While the business is not yet generating cash, it has successfully reduced its operating expenses through improved laboratory efficiency and the strategic use of AI to lower platform costs.
The balance sheet is strong with $665.2 million in cash and minimal debt. This cash pile provides a "runway" that management expects to last into early 2028 at current spending levels. Having this much cash is vital because it allows Recursion to continue its research and clinical trials without being forced to borrow money or sell new shares of stock during periods of market volatility.
Recursion is a well-funded research engine that has enough cash to reach several major clinical milestones before needing more capital.
Cash efficiency is improving as the company reduces its quarterly burn rate while advancing more drugs into human trials. Net cash used in operations fell from $132 million to $81.1 million year-over-year, proving that the company can scale its research without costs spiraling out of control.
Revenue from partners can be unpredictable and may stall if the platform does not produce enough high-quality drug leads for Roche and Bayer. A long gap between milestone payments would force the company to dip faster into its $665 million cash reserve, shortening its time to survive without new funding.
The "TechBio" industry is roughly $5 billion today and is projected to exceed $20 billion by 2028 as pharmaceutical companies move away from manual research. Pricing power in this industry is driven by the rarity and quality of biological data, which is the "fuel" for AI models. Recursion is a leading player in this space because it generates its own data in-house rather than relying on public records. This gives it a significant head start in training models that can actually predict whether a drug will work in humans.
The market for AI drug discovery is highly competitive and fragmented, but the barriers to entry are becoming massive due to the cost of computing power. The primary battle is no longer about who has the best code, but who has the most high-quality, proprietary biological data.
Schrodinger is the most established rival, but it focuses on chemistry and physics modeling rather than the full-scale biological imaging Recursion uses. AbCellera is the most dangerous threat because it has already proven its platform can discover a billion-dollar drug (Bamlanivimab for COVID-19) and has built deep relationships with Eli Lilly. Other players like Insilico Medicine are moving fast in China, potentially reaching the market with lower-cost AI-designed molecules first.
Recursion is holding its ground by building a "full-stack" operation that owns everything from the supercomputer to the wet lab. Its acquisition of Exscientia in 2024 further strengthened its chemical design capabilities, making it one of the few companies that can handle both the biology and the chemistry of a new drug.
Recursion’s primary source of protection is its intangible assets, specifically its massive, proprietary "BioHive" dataset of biological images and its NVIDIA-powered supercomputing core. The moat exists because a competitor would have to spend hundreds of millions of dollars and several years running millions of physical experiments to recreate Recursion’s data library. The company owns 504 NVIDIA H100 GPUs, giving it the compute power to process this data faster than any other biotech firm.
The financial metrics currently show a company in a heavy investment phase, with -47% ROIC and negative margins. However, the $70 million in revenue and 43% reduction in polyp burden in REC-4881 patients suggest that the intangible advantage is starting to produce real-world results. A real moat in biotech is only fully proven when a drug reaches the market, but the deep integration with Roche suggests high switching costs are forming.
The moat is strengthening as every new experiment adds data to the BioHive, making the "map of biology" more accurate for future searches.
Advanced REC-1245 from discovery to clinical trial in 18 months.
Maintained $665M cash while reducing operating cash burn by 38%.
Founder Chris Gibson remains Chairman with a significant personal stake.
Capital Allocation Track Record
Management has shown exceptional strategic judgment by pivoting Recursion from a pure research platform into a clinical-stage drug developer while maintaining a strong cash position. CEO Najat Khan, who previously led R&D strategy at Johnson & Johnson, brings the big-pharma discipline needed to navigate the FDA. The team’s ability to move drug candidates into human trials in under two years—twice as fast as the industry average—is clear evidence of a high-caliber leadership team that understands how to use technology to solve biological bottlenecks.
The primary risk is the company’s heavy dependence on a few key scientific leaders, though the board has successfully built a deep bench of pharma veterans. While the "TechBio" vision is central to the thesis, the execution is increasingly handled by experienced clinical officers like Vicki Goodman and Ben Mabey. The dual-class share structure keeps control with the founders and early backers, which protects the long-term vision but limits the ability of outside shareholders to force a change if clinical results stall.
We expect revenue to grow from $0.1B in FY2026 to $0.8B in FY2031 (~69% CAGR), with EPS growing from $-0.92 to $0.33. Revenue accelerates as the Recursion OS platform transitions from early research milestones to larger commercial payments and royalties from partners like Roche and Bayer. Automated laboratory workflows and AI-driven lead optimization allow the company to scale its drug pipeline without a linear increase in research headcount or physical infrastructure. EPS grows faster Operating margin expected to reach ~30% by FY2031.
Lead drug REC-4881 reaches the market for rare polyps. Success in this niche market would provide Recursion's first commercial revenue and prove the platform's validity.
Roche or Bayer expand partnerships to include more disease areas. Deeper integration with big pharma would provide hundreds of millions in additional milestone payments and royalties.
The Recursion OS identifies a blockbuster drug for a major cancer. Using AI to solve complex cancers would transform Recursion from a niche player into a global pharmaceutical giant.
Clinical trials for REC-4881 fail to meet FDA standards. If the lead drug fails, it would suggest the AI platform is better at finding "leads" than finding real cures.
Cash burn accelerates before the company reaches commercial sales. Running out of cash before 2028 would force a dilutive stock sale at a potentially low share price.
Competitors with larger datasets or better algorithms displace the Recursion OS. If a rival like AbCellera or a tech giant builds a superior biological map, Recursion's platform value would collapse.
Below is our estimate of current and future fair value, with detailed reasoning and assumptions. Fair value is a judgment, not a fact, and other analysts will likely land on different numbers. Use it as one data point in your research, and apply your own discretion in any investing decision.
We use a 5-year Discounted Cash Flow (DCF) model anchored to the company's projected 2031 profit inflection. This framework is the institutional standard for clinical-stage biotechs because current revenue is negligible and volatile; the real value lies in the probability-weighted future earnings of a platform that can generate multiple "shots on goal" simultaneously.
Applying a 30x terminal multiple to the projected FY2031 EPS of $0.33, then discounting at 10% and adjusting for the intervening cash burn, results in a $3.00 fair value. The 30x multiple sits at the upper end of mature biotech (15-25x) and reflects a premium for the NVIDIA-powered AI data moat, positioning Recursion closer to software-enabled peers like Schrödinger (SDGR, ~10x P/S) than traditional single-drug biotechs. This calculation follows the deterministic engine's projection for long-term earnings while penalizing for the ~$1.6B in cumulative burn required to reach that inflection point.
A Price-to-Sales (P/S) cross-check using the current 21.4x multiple on projected NTM revenue of $75M produces a fair value of $3.02, confirming our $3.00 DCF result. This multiple is consistent with Recursion's 4-year historical average and reflects the "platform premium" the market awards to the Roche/Genentech alliance, which has already yielded over $500M in milestones. This cross-check result is within 1% of our primary model, providing high conviction in the fair value.
We're assuming Recursion's cash runway remains intact into early 2028 as guided by management. This is supported by the $665 million cash balance and a projected annual operational burn of $390 million, which provides sufficient buffer to reach major clinical data readouts without needing immediate, dilutive equity financing.
We're assuming the company's FY2031 profitability target of $0.33 per share is achievable. This requires the "industrialization" of drug discovery to yield at least two successful commercial-stage products and a steady stream of high-margin milestone payments from the Roche and Bayer collaborations over the next five years.
We're assuming the NVIDIA partnership sustains Recursion’s infrastructure advantage via the BioHive-2 supercomputer. By treating biology as a "search problem" rather than a hypothesis problem, Recursion aims to identify more viable drug candidates at lower costs; we assume this infrastructure keeps R&D expenses from scaling linearly with the pipeline.
The biggest risk is a clinical failure or negative regulatory update for lead candidate REC-4881 in the second half of 2026. Such a result would not only delay commercialization but would also damage the "platform validation" thesis, likely knocking the fair value down to the company's liquidation/cash value of approximately $1.50 per share. Watch for any Phase 2 trial enrollment delays as an early signal of clinical friction.
Bear case ($1): REC-4881 Phase 2 data fails to show statistical significance, invalidating the predictive platform's "search" capability; or Strategic partners Roche or Bayer terminate a major collaboration program, cutting the cash runway significantly.
Bull case ($7): FDA grants Accelerated Approval for REC-4881 in late 2026, transitioning the company from research to commercial-stage; or Nvidia increases its equity stake or announces a direct licensing agreement for BioHive-2 technology to third-party biotechs.
Clearthesis wrote this report from 30 sources, including SEC filings, industry research, and recent news.
How did you like this thesis?
Your feedback helps us make reports better for you
© 2026 Clearthesis.ai · Report generated on June 23, 2026
This is an AI-generated analysis for informational purposes only and does not constitute financial advice. Data and analysis may not reflect recent developments if viewed significantly after the generation date. Always conduct your own due diligence before making any investment decisions.
The market is leaning bullish because Recursion uses automated computer power to replace the slow, hit-or-miss guessing game of traditional drug discovery. The company uses its BioHive-2 supercomputer to conduct millions of digital experiments weekly. This systematic approach aims to turn medicine development into a repeatable engineering process rather than a scientific gamble.
Skeptics think that the company remains a risky bet because it has yet to prove these automated patterns consistently lead to successful human clinical outcomes. While the platform can identify many potential drug targets quickly, the high failure rate of actual clinical trials means the technology still faces the same biological hurdles as traditional drug research.