Introduction: Performance fatigability is substantially greater when exercising in the severe-versus heavy-intensity domain. However, the relevance of the boundary between moderate-and heavy-intensity exercise, the gas exchange threshold (GET), to performance fatigability is unclear. This study compared alterations in neuromuscular function during work-matched exercise above and below the GET. Methods: Seventeen male participants completed work-matched cycling for 90, 110, and 140 min at 110%, 90%, and 70% of the GET, respectively. Knee extensor isometric maximal voluntary contraction (MVC), high-frequency doublets (Db100), low-to high-frequency doublet ratio (Db10:100), and voluntary activation were measured at baseline, 25%, 50%, 75%, and 100% of task completion. During the initial baseline visit and after each constant work rate bout, ramp-incremental exercise was performed, and peak power output and oxygen uptake ( VO 2peak ) were determined. Results: After the 70% and 90% GET trials, similar reductions in MVC (−14% ± 6% and −14% ± 8%, respectively, P = 0.175) and Db100 (−7% ± 9% and −6% ± 9%, respectively, P = 0.431) were observed. However, for a given amount of work completed, reductions in MVC (−25% ± 15%, P = 0.008) and Db100 (−12% ± 8%, P = 0.029) were up to 2.6-fold greater during the 110% than the 90% GET trial. Peak power output and VO 2peak during ramp-incremental exercise were reduced by 7.0% ± 11.3% and 6.5% ± 9.3%, respectively, after the 110% GET trial relative to the baseline ramp (P ≤ 0.015), with no changes after the moderate-intensity trials (P ≥ 0.078). Conclusions: The lack of difference in fatigability between the trials at 70% and 90% GET, coupled with the greater fatigability at 110% relative to 90% GET, shows that exceeding the moderate-to heavy-intensity boundary has implications for performance fatigability, while also impairing maximal exercise performance capacity.