Interview, Fireside Chat
How Autonomous Labs Will Transform Scientific Research: Ginkgo Bioworks’ Jason Kelly
- Biotechnology and biopharma industries face fundamental disruption where AI reasoning models paired with autonomous labs are expected to outperform human scientists by running experiments 24/7, iterating hypotheses daily, and leveraging data sharing among hundreds of AI agents.
- The technology aims to achieve a 10x increase in data per dollar by shifting costs from human overhead to reagents and increasing equipment utilization from under 20% to roughly 70% within centralized, smaller, and denser laboratories.
- Ginkgo Bioworks has shifted its two-year focus to optimizing the "back end" of engineering by solving technical hurdles in liquid handling and equipment integration, moving from manual bench work to a system where samples move on tracks between robotic arms rather than relying on humanoids.
- Automation infrastructure is expected to be sold as a subscription service with a future expansion into "automation-friendly reagents," while the broader life science tools industry is predicted to provide only incremental changes rather than shifting scientific fundamentals.
- Current biotech R&D strategies relying on manual labor and indirect rates are viewed as inefficiencies that will prompt re-evaluations of funding models by the NIH and biopharma heads, with costs expected to fall while distribution costs rise.
- Training data generated by autonomous labs is projected to create a feedback loop that could 10x or 100x the speed of scientific discovery, potentially allowing millions of people to become scientists by democratizing access to cloud lab services.
- Project Genesis, a Department of Energy initiative, is expected to double the acceleration of science in the "next few years" by integrating AI models into national labs, while the U.S. risks falling behind China in drug discovery innovation without widespread AI adoption.
- The biotech ecosystem is transitioning from isolated, low-utilization labs to centralized platforms utilizing "cloud code" that allow users to execute protocols via natural language without needing Python coding skills.
- Fundamental research is predicted to regain value as industries shift from engineering incremental improvements to breakthrough discoveries driven by the emergence of "super intelligent" AI in specific categories within months or years.
- Application spaces are expected to expand beyond the current 85% therapeutic market share to include consumer health, longevity, and molecular monitoring, though regulatory frameworks currently hinder non-disease treatments.
- ACTG-native models like Ark's Evo will evolve into powerful tools for accessing protein designs and synthesizing reagents, supporting reasoning models rather than replacing core engines.
- Key risks include the potential for getting "technologically surprised" by rivals, which poses a significant threat to national security, as well as regulatory barriers preventing the expansion of biotech applications outside traditional disease treatments.
- The pace of drug discovery is expected to accelerate significantly once manual experimental components are removed, addressing the trend of rising costs over the last 25 years driven by labor rather than a lack of tools.