AI Drug Discovery Beneficiaries
Richard@rich
−3.9%
since Sep 29
The idea
My thesis is that as AI pushes deeper into biology, the gains flow to two groups: the genomics and lab tool makers supplying the data these models need, and the large pharma companies with the pipelines and commercial muscle to actually bring AI-assisted drugs to market. I split the basket evenly between those two halves, add a layer of wet-lab bottleneck names like CROs and lab automation that any AI model still depends on, and keep a small sleeve of AI-native biotechs for asymmetric upside.
What it owns
- Established drugmakers with the infrastructure to run trials, navigate regulators, and commercialize drugs at scale, now deploying AI to compress discovery timelines. · Eli LillyIt operates one of pharma's largest pipelines and commercial organizations while investing heavily in AI discovery to accelerate candidate generation and time to market., AstraZenecaIt has embedded AI into discovery workflows and maintains the regulatory and sales infrastructure to bring AI-assisted candidates through to patients., Novo NordiskIt generates cash flow sufficient to fund AI-native drug discovery at scale while retaining the commercial machinery to launch and market what those programs produce., MerckIt has formed multiple partnerships to source AI-discovered candidates and possesses decades of expertise converting early-stage compounds into approved medicines., SanofiIt has publicly committed to building its pipeline through AI-first discovery methods, combining the AI innovation model with traditional pharma commercialization capability.32%
- Instrument and consumables suppliers whose sales accelerate when AI-driven research programs demand more experimental biological data to train and validate models. · Thermo Fisher ScientificIt manufactures the sequencers, instruments, and consumables that labs need to generate the experimental data AI models require for training and validation., DanaherIts bioprocessing and diagnostics units supply the equipment and services that produce the biological datasets AI-driven research depends on to move programs toward the clinic., IlluminaIt operates the majority of sequencing capacity worldwide, which becomes more critical if AI models demand increasing volumes of genomic data for research and drug development., Agilent TechnologiesIts analytical instruments are widely used in labs generating the research data that AI-driven drug discovery consumes in larger volumes as the field expands., 10x GenomicsIt produces single-cell analysis tools that deliver the granular biological data AI models are most effective at learning from and using for drug design., Twist BioscienceIt manufactures synthetic DNA for testing AI-predicted molecules in living systems, turning computational designs into experimental validation.32%
- Contract research organizations and lab automation vendors that perform the physical experiments AI models cannot replace, constraining how fast AI candidates can advance. · Charles River LaboratoriesIt conducts preclinical testing for most AI-designed drug candidates before they enter human trials, making it essential to converting AI predictions into validated compounds., BrukerIts structural biology instruments generate the protein and molecular data that train AI models and validate their predictions in real laboratory conditions., RevvityIts proteomics and screening platforms perform validation work that AI models cannot yet replace, sitting at the critical junction between computational design and wet-lab proof., Medpace HoldingsIt conducts clinical trials for smaller biotechs developing AI-discovered drugs, scaling with the wave of candidates entering human testing over the next years.17%
- Smaller companies built around AI-first drug discovery, where a single successful trial readout can drive rapid value creation over a concentrated timeframe. · Recursion PharmaceuticalsIt is built around AI-first drug discovery with its own pipeline, offering concentrated upside if any program reaches a positive trial readout within the next couple of years., SchrödingerIts physics-based computational platform is licensed across pharma for drug design and it operates its own pipeline, generating dual revenue streams from the AI-to-drug trend., Tempus AIIt aggregates and licenses the clinical and molecular datasets that AI drug-discovery programs need to train models and make predictions, directly monetizing the data supply.9%
- Cash10%