Faster or Better?AI Investment and Patent Time-to-Grant:Innovation Quality Upgrading as a Mechanism

Artificial intelligence (AI) is expected to accelerate innovation, yet its implications for the timing of value realization remain unclear. We theorize a quality–speed trade-off: AI investment enhances innovation quality by pushing firms toward more technologically sophisticated, frontier-oriented, and novel inventions, which increases examination complexity and slows patent granting. Using a panel of 5,273 Chinese listed firms (2012–2024), we measure firm AI investment from financial disclosures of AI-related intangible and fixed assets and estimate generalized estimating equation models. Results show that higher AI investment is associated with slower patent grant speed: a 10% increase in AI investment corresponds to a 3.39 percentage-point decrease in same-year grant probability and a 6.85 percentage-point decrease in grant probability within two years. Mechanism tests indicate that AI raises the share of invention patents, strengthens effects in high-tech and emerging domains, and increases both longitudinal and peer-relative text novelty. Regional data endowments amplify these effects, and the novelty gains are concentrated among the largest firms and are stronger for state-owned enterprises. Overall, the study reframes AI-enabled innovation as a quality–speed trade-off with regional data and firm heterogeneity as key boundary conditions.