−1.1%
since Oct 9
The idea
I believe AI models for biology are bottlenecked by a shortage of real experiment data, and that drug companies and AI labs will keep paying for it since nobody can run every experiment in-house. Rather than pick which AI company wins, this take owns the picks and shovels: lab equipment and consumables makers, the labs that generate and sell experiment data, and cell and tissue toolmakers tied to the shift away from animal testing. Most weight sits in steadier equipment suppliers, with smaller positions in younger data labs that can swing hard, plus some cash held back.
What it owns
- These companies sell the instruments, reagents and consumables that labs of every kind need to run experiments, so they benefit from research spending across many fields and customers. · Thermo Fisher ScientificIt sells the instruments, reagents and consumables that almost every lab buys, so it gets paid regardless of which research direction wins as long as lab spending stays healthy., DanaherIts bioprocessing, life science tools and diagnostics units sell recurring supplies and equipment to labs across many industries and research types., Agilent TechnologiesIt makes analytical instruments and lab consumables that are standard equipment in pharma and research labs, benefiting from steady instrument replacement and repeat supply orders., IlluminaIt produces sequencers that generate much of the world's genomic data, and every new dataset used to train biology models requires sequencing work., Waters CorporationIt makes liquid chromatography and mass spectrometry systems used to test drugs, especially newer biologic molecules, and provides recurring service and chemistry supply revenue., Twist BioscienceIt synthesizes DNA at scale, providing the raw material that many biology experiments and AI-designed tests require., RevvityIt sells reagents, instruments and software for life science research and diagnostics, serving academic and pharma customers who run ongoing experiments.46%
- These companies run outsourced experiments or store samples and generate datasets that pharma and biotech firms buy, turning lab work into a product sold to multiple customers. · Charles River LaboratoriesIt runs outsourced experiments for drug companies and is shifting toward non-animal testing as regulators move that way, positioning it to capture paid testing work as methods change., Recursion PharmaceuticalsIt runs automated labs that generate large biology and image datasets, then shares those with pharma and AI partners through research collaborations., Tempus AIIt packages clinical and molecular data into a product that drug companies license, building recurring revenue from patient sequencing and records combined with AI tools., Ginkgo BioworksIt operates large automated labs and sells lab services and datasets to outside customers, generating revenue as it scales up its automation., AzentaIt stores biological samples and runs gene sequencing and synthesis services that labs outsource, capturing recurring revenue from sample management and genomics work., AbsciIt pairs AI with its own wet lab to design antibodies and generate testing data, creating value through the combination of computation and experimental validation.28%
- These companies make the tools and materials that let researchers measure and work with individual cells and human tissue, the detailed data that modern biology research depends on. · Bio-TechneIt makes the proteins, antibodies and cell culture materials that human tissue models and organoids need to function, so demand grows as testing shifts toward human biology., 10x GenomicsIt makes single cell and spatial measurement tools that show what happens inside individual cells in real tissue, generating the detailed data that research increasingly demands., BrukerIts spatial biology tools and high-end mass spectrometry let labs measure tissue in fine detail, serving researchers who need rich biological data.17%
- Cash9%