Engineers at the University of Pennsylvania have developed LIBRIS, an automated robotic microfluidic platform designed to accelerate the creation of lipid nanoparticle (LNP) formulations. By generating approximately 1,000 distinct formulations per hour, the system operates 100 times faster than existing manual methods. This significant increase in speed aims to solve a critical data bottleneck that has hindered the integration of artificial intelligence into LNP design. LNPs serve as the essential delivery vehicles for mRNA therapies, including the widely used COVID-19 vaccines. Designing effective LNPs is complex because their performance depends on the specific ratios of multiple lipid components. While scientists can currently synthesize thousands of new ionizable lipids and test them, the actual mixing and formulation process remains the primary constraint. Traditional approaches, whether manual or automated, are too slow and often lack the consistency needed to build the massive, systematic datasets required to train predictive AI models. Without sufficient data, AI cannot identify the precise patterns linking chemical structures to biological outcomes. The new machine, short for Lipid nanoparticle Batch production via Robotically Integrated Screening, functions like a tiny factory. It uses tubes to feed different LNP components into a glass microfluidic chip encased in an aluminum housing. Inside the chip, microscopic channels mix these components under precisely controlled pressure. Unlike conventional systems that operate in serial fashion, LIBRIS features parallel channels capable of producing up to eight distinct formulations simultaneously. A robotic system moves plastic well plates beneath the chip to collect the resulting particle streams. Because the channels can be rapidly cleaned, the platform operates continuously, ensuring high throughput and batch-to-batch consistency. Michael J. Mitchell, an associate professor in Bioengineering, and David Issadore, a professor in the same field, lead the research team. They state that this technology could accelerate LNP development by as much as 100-fold. The design space for LNPs is vast, with potentially 10 to the power of 15 possible formulations, making manual exploration impossible. By rapidly generating well-defined libraries of particles, LIBRIS provides the necessary data volume for AI algorithms to learn and recognize effective structures. Currently, LNP discovery relies heavily on trial and error, where researchers test various combinations to see which perform best. The ultimate goal of this project is to shift from screening to rational design. With the data generated by LIBRIS, the team hopes to enable a future where researchers can specify desired particle properties in advance and use AI to determine the exact formulation required to achieve them. This transition promises to unlock the full potential of LNP-based therapeutics for treating antibiotic-resistant infections and genetic diseases, moving beyond the limitations of current empirical methods.