Ginkgo Cloud Lab Integrates GPT-5 for Biology

Ginkgo Bioworks' platform is not just a software layer; it's built on top of highly automated physical labs called "Foundries". These facilities, representing a half-billion-dollar investment in robotics, software, and genetic tools, are designed to scale the process of organism engineering. The company's goal is to make biology easier to engineer, transitioning from artisanal, manual lab work to a standardized, platform-based approach. The collaboration with OpenAI involved connecting GPT-5 to this cloud laboratory infrastructure in a closed-loop system. Over six months, the AI autonomously designed, executed, and analyzed over 36,000 experiments to optimize cell-free protein synthesis (CFPS). This process involved GPT-5 proposing experiments, robots carrying them out, and the results being fed back to the model to inform the next iterative cycle with minimal human input. This AI-driven approach yielded significant results, reducing the reaction component cost of producing a benchmark protein by 40% compared to the previous state-of-the-art. The system generated nearly 150,000 data points, and GPT-5 was even able to anticipate findings from the field, suggesting critical reagents that human experts had independently identified. Ginkgo is now commercializing the AI-improved reaction mix, demonstrating the direct business application of the research. Ginkgo's business model is twofold: it earns usage fees for its Foundry services and also takes a value share through royalties, milestones, or equity in the products its platform helps create. This horizontal platform approach allows them to serve diverse markets, including pharmaceuticals, agriculture, and industrial chemicals, rather than developing their own products. This move is part of a larger strategic pivot for Ginkgo, which recently announced plans to divest its biosecurity business to focus more on its core cell engineering services and investments in AI and automation. The company is betting that its massive repository of proprietary biological data, what it calls its "Codebase," will provide a key advantage in training specialized foundation models for biology. The broader market for AI in drug discovery is expanding rapidly, with analysts projecting the market to be worth $9 billion or more by the end of the decade. The integration of AI is seen as crucial for accelerating research, reducing costs, and designing novel proteins and therapeutics that would be difficult to find through traditional methods alone.