We report high-precision analyses of internally-normalised Ni isotope ratios in 12 bulk iron meteorites. Our measurements of 60Ni/61Ni, 62Ni/61Ni and 64Ni/61Ni normalised to 58Ni/61Ni and expressed in parts per ten thousand (‱) relative to NIST SRM 986 as ε60Ni5861,ε62Ni5861 and ε64Ni5861, vary by 0.146, 0.228 and 0.687, respectively. The precision on a typical analysis is 0.03‱, 0.05‱ and 0.08‱ for ε60Ni5861, ε62Ni5861 and ε64Ni5861, respectively, which is comparable to our sample reproducibility. We show that this ‘mass-independent’ Ni isotope variability cannot be ascribed to interferences, inaccurate correction of instrumental or natural mass-dependent fractionation, fractionation controlled by nuclear field shift effects, nor the influence of cosmic ray spallation. These results thus document the presence of mass-independent Ni isotopic heterogeneity in bulk meteoritic samples, as previously proposed by Regelous et al. (2008) (EPSL 272, 330–338), but our new analyses are more precise and include determination of 64Ni. Intriguingly, we find that terrestrial materials do not yield homogenous internally-normalised Ni isotope compositions, which, as pointed out by Young et al. (2002) (GCA 66, 1095–1104), may be the expected result of using the exponential (kinetic) law and atomic masses to normalise all fractionation processes. The certified Ni isotope reference material NIST SRM 986 defines zero in this study, while appropriate ratios for the bulk silicate Earth are given by the peridotites JP-1 and DTS-2 and, relative to NIST SRM 986, yield deviations in ε60Ni5861, ε62Ni5861 and ε64Ni5861 of −0.006‱, 0.036‱ and 0.119‱, respectively. There is a strong positive correlation between ε64Ni5861 and ε62Ni5861 in iron meteorites analyses, with a slope of 3.03±0.71. The variations of Ni isotope anomalies in iron meteorites are consistent with heterogeneous distribution of a nucleosynthetic component from a type Ia supernova into the proto-solar nebula.