JPMorgan routes nearly all of $320bn in 2035 robotaxi revenue through Tesla's own fleet

JPMorgan has robotaxi revenue at roughly $320 billion in 2035 ^[3] and total company revenue at about $203 billion in 2030 ^[5]. One line item, five years further out, is about 1.6 times everything the bank expects Tesla to sell in 2030 ^[15].

That line item only works if someone owns the cars. JPMorgan's stated reason for putting the fleet inside Tesla is vertical integration: the company builds the cars, writes the software and collects the fares ^[10]. The bank lists what central ownership buys, which is control of maintenance schedules, vehicle cleanliness, insurance costs and the customer experience end to end ^[9]. Tesla also keeps the fare, the software margin and the data ^[9]. Under that assumption the vehicles earning the fares are Tesla's own assets, and Tesla pays for them ^[18].

Around 9 million Tesla vehicles have contributed roughly 10 billion cumulative miles of Full Self-Driving data ^[8], an average of about 1,100 miles per car ^[16]. That installed base is what the peer-to-peer story needed. In the bank's model those cars are a training input, and the fleet capacity comes from Cybercab ^[7].

Zero owner-supplied revenue is JPMorgan's channel assumption ^[4], not a Tesla policy announcement, and a bank can put a channel at zero and be wrong about it. One thing would prove this reading wrong: Tesla opening owner-supplied cars in Austin, Dallas, Houston or the Bay Area ^[6] with a published take rate. Then 9 million vehicles already on the road become fleet capacity.

The rating stops short of underwriting the $320 billion. Cryptobriefing, which reported the note, says the $475 target implies meaningful appreciation while also implying the stock is not a screaming bargain at current levels ^[14]. Cryptobriefing did not report the price Tesla traded at before the upgrade, so the implied upside cannot be recomputed. JPMorgan flagged regulatory approval as a city-by-city, state-by-state grind, and safety validation for fully autonomous vehicles without human supervision as ongoing ^[11]. It also raised the question of whether a camera and AI stack can match the safety record of Waymo's lidar, radar and camera combination ^[12]. Waymo has run commercial robotaxi service for years, Amazon-backed Zoox is testing purpose-built vehicles, and Baidu's Apollo Go and Pony.ai are scaling in China ^[13].