A primary temperature scale requires realising a unit in terms of its definition. For high temperature radiation thermometry in terms of the International Temperature Scale of 1990 this means extrapolating from the signal measured at the freezing temperature of gold, silver or copper using Planck's radiation law. The difficulty in doing this means that primary scales above 1000 °C require specialist equipment and careful characterisation in order to achieve the extrapolation with sufficient accuracy. As such, maintenance of the scale at high temperatures is usually only practicable for National Metrology Institutes, and calibration laboratories have to rely on a scale calibrated against transfer standards. At lower temperatures it is practicable for an industrial calibration laboratory to have its own primary temperature scale, which reduces the number of steps between the primary scale and end user. Proposed changes to the SI that will introduce internationally accepted high temperature reference standards might make it practicable to have a primary high temperature scale in a calibration laboratory. In this study such a scale was established by calibrating radiation thermometers directly to high temperature reference standards. The possible reduction in uncertainty to an end user as a result of the reduced calibration chain was evaluated.