Purpose: To quantify the impact of a polarized distribution of training intensity on performance and fatigue in elite swimmers. Methods: Twenty-two elite junior swimmers (12 males, age = 17 [3] y, and 10 females, age = 17 [3] y) participated in a crossover intervention study over 28 wk involving 2- × 6-wk training periods separated by 6 wk. Swimmers were randomly assigned to a training group for the first period: polarized (81% in zone 1, blood lactate concentration, [La]b ≤ 2 mmol·L−1; 4% in zone 2, 2 mmol·L−1 < [La]b ≤ 4 mmol·L−1; and 15% in zone 3, [La]b > 4 mmol·L−1) or threshold (65%/25%/10%). Before and after each period, they performed a 100-m maximal swimming test to determine performance, maximal [La]b, and oxygen consumption and an incremental swimming test to determine speed corresponding to [La]b = 4 mmol·L−1 (V4 mmol·L−1). Self-reported indices of well-being were collected with a daily questionnaire. Results: Polarized training elicited small to moderately greater improvement than threshold training on 100-m performance (within-group change ± 90% confidence interval: 0.97% ± 1.02% vs 0.09% ± 0.94%, respectively) with less fatigue and better quality of recovery. There was no substantial gender effect. No clear differences were observed in physiological adaptations between groups. Conclusions: In elite junior swimmers, a 6-wk period of polarized training induced small improvements in 100-m time-trial performance and, in combination with less perceived fatigue, forms a viable option for coaches preparing such cohorts of swimmers for competition.