Abstract Revolutions in shared mobility services, vehicle electrification, and automated vehicle technology will affect urban traffic patterns, energy use and CO2 emissions, the automotive industry, public transportation, and more. This paper examines the monetary costs of these innovations for users in the near-term (approximately 2020) and how they may evolve in the long-term (approximately 2030–2035). We estimate traveler costs for light duty vehicle trips on a per-mile basis, and investigate their sensitivity to vehicle powertrain, vehicle size, travel mode and intensity of vehicle use, and DC charging assumptions. To highlight differences between human and automated driving, we consider only autonomous vehicle scenarios in the long-term. We document three main findings. First, as battery costs continue to drop over the next decade, private battery electric vehicles will become more cost-competitive with internal combustion vehicles; and in high-mileage ridesourcing applications, electric vehicles will be much more cost-competitive. Second, near-term ridesourcing trips will likely remain about 4–5 times the per-mile cost of driving one's own car, while pooled trips cut this factor to around 3. Third, in the long-term automated vehicles may make ridesourcing cheaper than driving one's own vehicle. Even if the manufacturing cost of automated vehicles remains high, this cost will be minor when amortized over a service life of 400,000 miles. These findings are unchanged even with significant variations in assumed future battery and automation costs, electricity (charging) cost, vehicle insurance and maintenance cost, and ridesourcing providers' overhead rates.

Highlights • As batteries become cheaper, private battery electric vehicles will become more cost-competitive with internal combustion vehicles. The full monetary cost of battery electric vehicle travel for households will be about equal to that of internal combustion engine vehicles within a few years, and already is for commercial (longer distance) vehicles. • Near-term (ca. 2020) ridesourcing trips, involving drivers, are about 4–5 times the per-mile cost of driving one's own car, while pooled trips cut this factor to around 3. • Assuming that in the long term (2030-2035), automated vehicles are available at a cost of $10,000 more than conventional driven vehicles, the net savings from eliminating the driver will make ridesourcing about 30% cheaper than driving one's own vehicle. Pooled trips will continue to provide savings but the cost advantage is likely to decline as overall costs decline. Policy and other implications.• For ridesourcing with EVs, the high mileage and associated energy savings could make these vehicles cost effective; the main question is whether range and recharging considerations create obstacles to their use as high mileage vehicles. • Ridesourcing in the near-term may fail competitive without private subsidies and suffer even more if regulations get severe (e.g. AB 5 in California, or Congestion Pricing in NYC). The question is whether public authorities could invest in bike infrastructure to facilitate micromobility providers' business, smooth capping on number of ebikes and scooters allowed per provider, facilitate POOL services, or support providers becoming multi-modal aggregator platforms. • The low cost of ridesourcing with driverless cars might lead to increased VMT. Policy interventions (such as pricing) could be considered to encourage more collective (POOL, High Capacity Vehicles) travel in the future coupled with multi-modal vision that includes micromobility.