We study an infinite-horizon, stochastic, dynamic optimization problem for an on-demand ride-hailing platform. The decisions include a one-time pricing decision per service at the beginning of the horizon, and real-time driver-customer-matching and driver-routing decisions during the planning horizon. The objective is to minimize the long-run average profit across time. We first establish a theoretical upper bound on the average profit under any policy. We then design a near optimal policy whose performance gap with respect to this upper bound vanishes in a meaningful asymptotic regime. To the best of our knowledge, this is the first policy in the ride-hailing literature with a theoretical performance guarantee for the actual spatial setting.