The plasma startup with ITER-relevant sub-mPa neutral pressures in the JT-60SA superconducting tokamak is studied using a zero-dimensional (0D) power balance model for the plasma burn-through phase. In the ohmic startup scenario assumed here, the lower limit of the operational neutral pressure is determined by the condition for the Townsend avalanche at a high limit, whereas the upper limit depends on the available ohmic heating power for the burn-through with the base power supply (PS) of the superconducting coils, corresponding to 0.2–0.3 V m−1. The operational conditions that lead to RE generation are evaluated to mitigate the risk of RE discharges. The simulation shows that startup REs are generated during the temperature rise immediately after the burn-through under relatively low-density conditions, and then the RE current value is gradually increased in balance with the RE loss processes on the timescale of Ip ramp-up. Therefore, it is expected that REs can be avoided or mitigated if the electron density during the early build-up phase is tuned by a combination of prefill gas and additional fuelling. However, a startup with an excessive prefill gas will lose the margin of the density control, and a full RE discharge can be triggered when the plasma burn-through fails owing to a high concentration of impurities such as oxygen, if the operational point is marginal to the burn-through limit. The present modelling work implies that preparing the operation with sufficient margins against the failure of impurity burn-through is important for mitigating the risk of unintended plasma termination. Although the carbon wall environment may have a lower risk of RE generation, to alleviate the operational uncertainties of the new device, strategies for expanding the operational window are also discussed in support of JT-60SA initial operation.