Integrating growing wind and solar generation while maintaining supply adequacy is a key challenge for sustainable power systems. However, the effects of different thermal retrofit pathways on renewable curtailment and modeled unserved energy remain unclear as renewable capacity expands. This study compares deep peak-shaving, fast-ramping, and combined retrofits using a PyPSA baseline and an 8760-h system dynamics simulation covering 31 Chinese provincial nodes. Doubling renewable capacity increases actual wind and solar generation by 51.4%, while the utilization rate falls from 91.4% to 69.2%. At baseline renewable capacity, the deep peak-shaving scenario S1 reduces annual curtailment by 25.0 TWh but produces 45.49 GWh of modeled unserved energy. S1 also reduces curtailment at renewable capacity multipliers of 1.50 and 2.00, whereas the fast-ramping scenario S2 has little effect on annual curtailment. At equal 50% retrofit coverage, the combined scenario S3 preserves S1’s curtailment benefit while nearly eliminating its supply shortfall. The expanded combined scenario S4, with 80% coverage, further reduces curtailment to 234.9 TWh at baseline renewable capacity. Considering curtailment and modeled unserved energy jointly, the combined retrofit provides the most favorable outcome among the equal-coverage pathways for supporting renewable integration while maintaining supply adequacy.