We investigated the effects of adaptation to a ketogenic low carbohydrate (CHO), high fat diet (LCHF) during 3 weeks of intensified training on metabolism and performance of world-class endurance athletes. We controlled three isoenergetic diets in elite race walkers: high CHO availability (g kg−1 day−1: 8.6 CHO, 2.1 protein, 1.2 fat) consumed before, during and after training (HCHO, n = 9); identical macronutrient intake, periodised within or between days to alternate between low and high CHO availability (PCHO, n = 10); LCHF (< 50 g day−1 CHO; 78% energy as fat; 2.1 g kg−1 day−1 protein; LCHF, n = 10). Post-intervention, Symbol during race walking increased in all groups (P < 0.001, 90% CI: 2.55, 5.20%). LCHF was associated with markedly increased rates of whole-body fat oxidation, attaining peak rates of 1.57 ± 0.32 g min−1 during 2 h of walking at ∼80% Symbol. However, LCHF also increased the oxygen (O2) cost of race walking at velocities relevant to real-life race performance: O2 uptake (expressed as a percentage of new Symbol) at a speed approximating 20 km race pace was reduced in HCHO and PCHO (90% CI: −7.047, −2.55 and −5.18, −0.86, respectively), but was maintained at pre-intervention levels in LCHF. HCHO and PCHO groups improved times for 10 km race walk: 6.6% (90% CI: 4.1, 9.1%) and 5.3% (3.4, 7.2%), with no improvement (−1.6% (−8.5, 5.3%)) for the LCHF group. In contrast to training with diets providing chronic or periodised high-CHO availability, and despite a significant improvement in Symbol, adaptation to the topical LCHF diet negated performance benefits in elite endurance athletes, in part due to reduced exercise economy.