Pando Bioscience

The pitch worth ignoring is the AI framing, because every enzyme company now claims it. What is worth reading is the screening throughput claim — 1000-fold more variants, 75% faster, 80% cheaper — because that is the actual bottleneck in directed evolution and it is the number that determines whether a custom program finishes in months or stalls. Pick Pando when you have an enzyme step that is eating your cost of goods, or a defined target with a measurable assay behind it. The small-molecule biocatalysis work and the Phase II DNA-modifying enzyme program for diagnostics are the clearest signal that this is scoped, milestone-driven work rather than a research collaboration someone keeps renewing. If all you need is amplification, skip the conversation entirely and buy PD-Phi29 or something from the Saiyan-Phi29 family. Off-the-shelf polymerases are the fastest path and Pando's own benchmarks — fidelity, chimera rate, processivity, reaction rate — are the ones to compare against whatever you run today. Pass if you have no internal assay capability. Pando's model improves because real wet-lab data goes back in, which means your team is on the hook for validation. Companies expecting a black box that hands over a finished enzyme will find the loop is a partnership, not a product. The closest alternative is doing directed evolution in-house, and for some targets that is still right. What in-house work rarely gives you is the compounding curve Pando reports: roughly 10x more validated hits in round two than round one, and activity gains that random mutagenesis does not replicate. Watch the infrastructure signals — SOC 2 Type II in progress, a Pando-DB at 22 billion annotated sequences, conference presence from CPHI Milan in October 2026 through ASHG Montreal and AMP