The manuscript “Gearbox Bearing Crack Growth Prognostics with Physics-Informed Machine Learning and Uncertainty Quantification” presents a solid and valuable study. The use of the X-TFC framework for predicting the remaining useful life of gearbox bearings is innovative, and the integration of fracture mechanics with physics-informed machine learning is carried out very well. Results on both vibration- and SCADA-based health indices are convincing, and the uncertainty analysis adds further strength.

The paper is original, timely, and clearly relevant for the wind energy and prognostics fields. The methods are well explained, the results are well supported, and the conclusions are logical. The writing is clear and well structured, making the paper accessible to a wide audience.

I do not see any major issues, and the manuscript is already at a very good standard.

Recommendation: Accept as is.

In the paper “Gearbox Bearing Crack Growth Prognostics with Physics-Informed Machine Learning and Uncertainty Quantification”, the authors introduce a novel physics-informed machine learning framework, the eXtreme Theory of Functional Connections (X-TFC), for analyzing and predicting the remaining useful life (RUL) of wind turbine gearbox bearings under fatigue crack growth. The key strength of this methodology lies in its ability to integrate both data-driven learning and physics-based knowledge into the neural network training process, thereby enhancing predictive performance.

In addition, the authors develop an uncertainty quantification approach based on Monte Carlo simulations, enabling the generation of confidence intervals for the RUL predictions. The framework is applied to both vibration-based and SCADA-derived health-index datasets. Results demonstrate strong predictive accuracy even under data-scarce conditions, along with low computational cost. Notably, the method achieves RUL errors below 60 hours even with limited training data, highlighting the effectiveness of physics-informed machine learning in improving neural networks inference.

The paper is well-structured, clearly presenting both the methodology and the results, and effectively demonstrates the potential of X-TFC for real-time bearing health monitoring. This work represents a valuable contribution to wind energy science and advances the state of the art in bearing health monitoring. I consider the topic highly relevant for the readers of WES journal and recommend the paper for publication without revisions, as no major concerns are identified.