© 2017 Author(s). We analyze the prospects for thermal spin-injection from iron into gallium arsenide via the application of electrical noise. By estimating the applied effective temperature-equivalent gradients, we characterize the magnitude of any electrical part of the thermal spin-injection efficiency or the spin-dependent Seebeck effect. The magnitude of the non-local spin signal associated with this effect suggests that temperature differences on the order of ∼100 K would be needed for true thermal spin-injection experiments. The large size of the effective temperature gradients induced by the noise-based method means that even very small thermo-electric effects can be quantified.