The first replicated productivity trials of the C4 perennial grass Miscanthus × giganteus in the United States showed this emerging ligno-cellulosic bioenergy feedstock to provide remarkably high annual yields. This covered the 5 years after planting, leaving it uncertain if this high productivity could be maintained in the absence of N fertilization. An expected, but until now unsubstantiated, benefit of both species was investment in roots and perennating rhizomes. This study examines for years 5–7 yields, biomass, C and N in shoots, roots, and rhizomes. The mean peak shoot biomass for M. × giganteus in years 5–7 was 46.5 t ha−1 in October, declining to 38.1 t ha−1 on completion of senescence and at harvest in December, and 20.7 t ha−1 declining to 11.3 t ha−1 for Panicum virgatum. There was no evidence of decline in annual yield with age. Mean rhizome biomass was significantly higher in M. × giganteus at 21.5 t ha−1 compared to 7.2 t ha−1 for P. virgatum, whereas root biomass was similar at 5.6–5.9 t ha−1. M. × giganteus shoots contained 339 kg ha−1 N in August, declining to 193 kg ha−1 in December, compared to 168 and 58 kg ha−1 for P. virgatum. The results suggest substantial remobilization of N to roots and rhizomes, yet still a substantial loss with December harvests. The shoot and rhizome biomass increase of 33.6 t ha−1 during the 2-month period between June and August for M. × giganteus corresponds to a solar energy conversion of 4.4% of solar energy into biomass, one of the highest recorded and confirming the remarkable productivity potential of this plant.