Lab Bottleneck Picks and Shovels
Henry Hirshland@henry
−2.8%
since Oct 6
The idea
I believe AI will keep accelerating biomedical research, but every new hypothesis still has to be tested at the bench, so the real bottleneck shifts to the companies supplying what research consumes. This basket spreads across that lab stack: consumables and bioprocessing, instruments and automation, sequencing and biological data, research software, and contract research, with a small broad biotech sleeve for upside without picking individual drug winners. It holds some cash and avoids concentrated single-name bets, built as a five-plus year structural hold.
What it owns
- Reagents and single-use supplies get consumed by every experiment, so rising research volume drives sales regardless of which therapies succeed. · DanaherIt manufactures the bioprocessing equipment and filtration systems that every biologic drug production run depends on., Thermo Fisher ScientificIt supplies the reagents and consumables that laboratories reorder continuously, making its revenue insensitive to which individual drugs succeed., RepligenIt produces the consumable components used in biologic drug production, earning revenue based on manufacturing volume rather than specific drug outcomes., AvantorIt distributes the everyday laboratory chemicals and supplies that get consumed continuously as research activity increases., West Pharmaceutical ServicesIt manufactures vials, stoppers, and injection components required for every injectable therapy approved, benefiting from a wider pipeline of new drugs.28%
- More research hypotheses require more lab capacity, pushing demand for the analytical instruments and automation systems that labs must expand or upgrade. · Agilent TechnologiesIt sells the analytical instruments laboratories purchase when expanding capacity to handle increased experimental workloads., Waters CorporationIt manufactures chromatography and mass spectrometry systems required to verify sample composition, work that must happen physically and cannot be automated away., Mettler-Toledo InternationalIt makes precision weighing and measurement instruments found on nearly every lab bench, benefiting from lab expansion and rising experiment volumes., RevvityIt provides discovery-stage tools and reagents consumed when researchers test hypotheses generated by computational models., BrukerIt manufactures high-end instruments like NMR and advanced mass spectrometry systems used in structural biology and deep research., AzentaIt sells sample automation and storage systems, the physical infrastructure labs need when experiment volumes explode beyond manual capacity.24%
- AI models depend on the quality and volume of biological data used to train them, creating steady demand for sequencing and sample preparation. · IlluminaIt operates on a per-run revenue model for sequencing, benefiting directly from rising research volume and AI training data demand., QiagenIt supplies sample preparation kits required before sequencing or molecular testing, collecting revenue from every test performed., Twist BioscienceIt synthesizes custom DNA sequences, providing the molecular building blocks for testing compounds designed through AI-assisted drug discovery., 10x GenomicsIt sells single-cell sequencing tools that generate the fine-grained biological data needed to train advanced AI models.15%
- The computational systems that design experiments and run AI models represent the digital layer enabling faster research workflows. · NVIDIAIt provides the compute hardware on which biomedical AI models train and run, forming the computational foundation for accelerated research., SchrödingerIt provides molecular simulation software that narrows which compounds are worth testing experimentally, supporting AI-assisted drug design workflows., Veeva SystemsIt operates the software systems pharmaceutical companies use to manage research and trials, benefiting from higher research throughput and faster workflows.11%
- When drug developers generate more candidates than they can test internally, they outsource to specialized contract research organizations. · IQVIA HoldingsIt conducts clinical trials and manages health data, seeing higher demand as AI-generated candidates advance toward human testing., Charles River LaboratoriesIt offers preclinical testing services that every drug candidate must pass through, creating demand growth when AI expands the pipeline of compounds to test.8%
- A diversified basket of drug makers captures the value of AI-enabled therapies without betting on individual companies. · SPDR S&P Biotech ETFIt holds a diversified portfolio of small biotech companies, capturing upside from AI-enabled therapies without predicting individual winners.6%
- Other · TEM2%
- Cash6%