Accelerating Reaction Discovery through AI-HTE Integration: Nitrene-Mediated C–O Coupling as a Validated Case Study

Carbamates constitute ubiquitous structural motifs across pharmaceuticals, agrochemicals, and materials, yet isocyanate-free strategies that accommodate sensitive functional groups remain underdeveloped. We report an efficient ruthenium-catalyzed C–O coupling of dioxazolones with phenols and alcohols that delivers carbamates under mild conditions with CO2 as the sole byproduct. This transformation establishes dioxazolone-derived nitrenes as a general platform for carbamate synthesis, exhibiting a broad substrate scope, functional-group tolerance, and scalability. Reaction development employed a preregistered, high-throughput workflow integrating machine learning. A model trained on phenolic substrates prospectively predicted reaction yields for unseen alcohol substrates, enabling direct prioritization and translation from microscale screening to preparative synthesis without substrate-specific reoptimization. This approach facilitated the late-stage functionalization of bioactive molecules and gram-scale preparation. Mechanistic studies and DFT support a novel Ru–nitrenoid pathway with Curtius-type rearrangement, where C–O bond formation is rate-determining. This work establishes both a practical catalytic platform for carbamate synthesis and a prospective strategy for integrating data-driven tools into the synthetic methodology.