Arsenobetaine has always been referred to as a non-toxic but readily bioavailable compound and the available data would suggest that it is neither metabolised by nor accumulated in humans. Here this study investigates the urine of five volunteers on an arsenobetaine exclusive diet for twelve days and shows that arsenobetaine was consistently excreted by three of the five volunteers. From the expected elimination pattern of arsenobetaine in rodents, no significant amount of arsenobetaine should have been detectable after 5 days of the trial period. The arsenobetaine concentration found in the urine was constant after 5 days and varied between 0.2 and 12.2 µg As per L for three of the volunteers. Contrary to the established belief that arsenobetaine is neither accumulated nor generated by humans, the presented results would suggest that either accumulated arsenobetaine in the tissues is slowly released over time or that arsenobetaine is a human metabolite of dimethylarsinic acid or inorganic arsenic from the trial food, or both. Either possibility is intriguing and raises fundamental questions about human arsenic metabolism and the toxicological and environmental inertness of arsenobetaine.