Intrinsic toroidal rotation of the deuterium main-ions in the core of the DIII-D tokamak is observed to transition from flat to hollow, forming an off-axis peak, above a threshold level of direct electron heating. Nonlinear gyrokinetic simulations show that the residual stress associated with electrostatic ITG turbulence possesses the correct radial location and stress structure to cause the observed hollow rotation profile. Residual stress momentum flux in the gyrokinetic simulations is balanced by turbulent momentum diffusion, with negligible contributions from turbulent pinch. Prediction of the velocity profile by integrating the momentum balance equation produces a rotation profile that qualitatively and quantitatively agrees with the measured main-ion profile, demonstrating that fluctuation-induced residual stress can drive the observed intrinsic velocity profile. Introduction-Turbulent transport in fluid systems such as Earth's atmosphere, stellar and laboratory plasmas can produce striking, self-organized features in the energy, density and momentum[1, 2] of the medium. In a tokamak, the phenomena of self-organized angular momentum creates an "intrinsic" rotation, where differential fluid flow can arise spontaneously. Intrinsic rotation here is defined by the magnitude and shape of the toroidal angular velocity profile that self-organizes in the absence of auxiliary torque injection, and can exhibit a wide range of nonlinear phenomenology[3] including threshold behavior across subtle changes in plasma conditions, as well as bifurcations in the rotation direction [4,5]. This plasma rotation is well known to have beneficial effects on energy confinement[6] and plasma stability [7]. In future large tokamaks such as ITER the rotation profile of the main-ions is expected to be largely determined by intrinsic processes because the ability of auxiliary torque to drive plasma rotation will be much smaller than in existing devices. ITER will operate in a regime that is dominantly electron heated by fusion alpha particles, where the heating of the ions will be dominantly through collisional energy exchange, motivating intrinsic rotation studies and model validation with direct electron heating and nearly equilibrated electron and ion temperatures. These considerations motivate achieving a validated predictive capability for main-ion intrinsic toroidal rotation for projection to ITER.