Researchers reported an AI-guided approach to engineer thermostable, solid-state mRNA vaccines designed to avoid ultra-cold storage constraints. Using an AGENT framework that couples high-throughput experimentation with Bayesian optimization, the study reported formulations retaining full bioactivity after storage at 37°C for more than two months and producing immune responses non-inferior to fresh vaccines in rodents and nonhuman primates. The work also described cold-chain-free delivery via microneedle patches in preclinical models, expanding potential use cases for mRNA beyond logistics-limited settings. If replicated in broader translational work, the formulation strategy could influence future vaccine platform design for speed and global deployment.