AI Biology Data Loop
Victor Angel@vanmosti
−3.3%
since Oct 9
The idea
I believe AI labs pushing into biology are turning wet-lab data into the scarce resource, since hypotheses are now cheap but cause-and-effect experiments aren't. This take owns the whole loop that produces that data: instruments that measure cells, the synthetic DNA and building blocks every design starts from, the reagents each run consumes, and the automated labs and outsourced testing that turn designs into results. More weight sits in smaller, higher-growth names where AI-driven demand already shows up, balanced by steadier instrument makers paid per experiment regardless of who wins.
What it owns
- Sequencing, single-cell analysis, spatial mapping and protein measurement turn cells into data that AI models use to learn cause and effect. · Illumina, Inc.It gets paid each time cells are sequenced, and drug companies are using its tools to train virtual cell models on large datasets., 10x Genomics, Inc.It makes instruments that read individual cells and map their location in tissue, providing the cause-and-effect data that virtual cell models need to learn., Bruker CorporationIts mass spectrometers measure proteins, a layer of biology that DNA data alone cannot show, and recently mapped a draft atlas of human blood proteins., Quanterix CorporationIt measures tiny amounts of protein in blood and expanded its biomarker partnerships, so it earns its place if AI-driven biomarker work grows its sales., Pacific Biosciences of California, Inc.Its long-read sequencing catches genetic details that short reads miss, and AI partnerships are using its data to train models for rare disease RNA drugs., Waters CorporationIts instruments check what is actually in a sample, a step every drug candidate goes through, and it partners with others on companion diagnostics.33%
- Every experiment consumes reagents, so these suppliers get paid per run regardless of which lab or model becomes dominant. · Bio-Techne CorporationIt sells the proteins, antibodies and other reagents that get used up in almost every cell experiment., Thermo Fisher Scientific Inc.It sells the consumables and outsourced services that most labs run on, so it gets paid per experiment regardless of which AI approach wins., Danaher CorporationIt is launching an AI-powered autonomous lab that runs the design, make, test and learn cycle much faster and includes its reagents and tools in the platform., Revvity, Inc.It sells reagents and screening tools used to test thousands of compounds or gene edits at once, the scale that AI-designed work requires., Agilent Technologies, Inc.It makes lab instruments and oligonucleotides, the short DNA and RNA strands used in experiments, so it touches both building blocks and measurement.25%
- AI-designed proteins and gene edits start as synthetic DNA or nucleic acids, so these companies see AI order volume before the rest of the pipeline. · Twist Bioscience CorporationIt synthesizes the DNA that AI-designed experiments begin with and expects AI drug discovery orders to grow rapidly in the coming year., Maravai LifeSciences Holdings, Inc.It makes the nucleic acid building blocks and enzymes that DNA and mRNA workflows depend on., Codexis, Inc.It is developing enzymatic DNA synthesis as an alternative to traditional chemistry, which could ease bottlenecks if that process wins paying customers.16%
- Robots, sample handling and outsourced testing provide the hands that turn AI designs into validated results at scale. · Azenta, Inc.It stores and handles biological samples and runs outsourced sequencing and gene synthesis services for research customers., Ginkgo Bioworks Holdings, Inc.It operates robot-driven labs and sells its automation platform, providing outsourced capacity for the design and testing loops that AI drives., Charles River Laboratories International, Inc.It runs the preclinical safety and efficacy tests that every drug candidate, including AI-designed ones, must pass before human trials.15%
- Cash11%