Presolar grains are stardust particles that condensed in the ejecta or in the outflows of dying stars and can today be extracted from meteorites. They recorded the nucleosynthetic fingerprint of their parent stars and thus serve as valuable probes of these astrophysical sites. The most common types of presolar silicon carbide grains (called mainstream SiC grains) condensed in the outflows of asymptotic giant branch stars. Their measured silicon isotopic abundances are not significantly influenced by nucleosynthesis within the parent star but rather represent the pristine stellar composition. Silicon isotopes can thus be used as a proxy for galactic chemical evolution (GCE). However, the measured correlation of 29Si/28Si versus 30Si/28Si does not agree with any current chemical evolution model. Here, we use a Monte Carlo model to vary nuclear reaction rates within their theoretical or experimental uncertainties and process them through stellar nucleosynthesis and GCE models to study the variation of silicon isotope abundances based on these nuclear reaction rate uncertainties. We find that these uncertainties can indeed be responsible for the discrepancy between measurements and models and that the slope of the silicon isotope correlation line measured in mainstream SiC grains agrees with chemical evolution models within the nuclear reaction rate uncertainties. Our result highlights the importance of future precision reaction rate measurements for resolving the apparent data–model discrepancy.