Marine dinitrogen (N2) fixation is quantitatively the most important source of new nitrogen (N) to the ocean (Duce et al., 2007). While marine nitrogen inventory has been balanced for the past 3000 years (Codispoti, 2005) marine nitrogen budget could not be balanced when based on direct rate measurements. Nitrogen loss has been calculated to exceed the gain from N2‐fixation by approximately 200 TgN/a (Codispoti, 2005; Karl et al., 2002). While the possibility of N‐imbalance cannot be fully excluded at this point, N2‐fixation has most likely been significantly underestimated in the past for two major reasons. Firstly, it was recently demonstrated that a methodological problem is associated with the commonly used 15N2‐tracer technique and the respective calculation (Montoya et al., 1996), which has significantly underestimated N2‐fixation rates. Consequently, a revised method was established (Mohr et al., 2010) and its application in the Atlantic found up to 6‐fold higher N2‐fixation rates compared to those determined with the classical method (Großkopf et al., 2012). Secondly, N2‐fixation was conventionally thought to be most active in nutrient‐depleted waters such as subtropical gyre centers, but a recent modeling study suggested a close spatial link between N‐loss in oxygen minimum zones (OMZs) and N2‐fixation in the adjacent surface ocean (Deutsch et al., 2003). Very recently, we showed the prevalence of the functional key gene for N2‐fixation (nifH) – belonging to a set of novel clusters of diazotrophs – present in ecological niches that have previously been underestimated (Löscher et al., 2013). Thus, we hypothesize that application of N2‐fixation measurements using the modified method combined with the consideration of novel habitats and novel diazotrophic clusters will significantly reduce the imbalance in the oceanic fixed‐nitrogen budget. Our results are of global importance as the availability of fixed N controls primary production in large areas of the surface ocean.