Let's dimension the US robotaxi market (since market participants seem unwilling to do so). People pattern match against structurally ~$3 per mile point to point mobility products and so misunderstand the potential scope of robotaxi as it becomes mass accessible. The average US adult spends nearly an hour per day driving. The imputed labor cost of all that manual piloting runs in excess of $4 trillion per year. In addition we pay $1.6 trillion annually for the actual service of driving point to point. By giving people back time (for which they don't have to pay full freight) and winning spend share, we think the US market could approach $4 trillion annually at saturation. Given reasonable expectations of supply diffusion and consumer adoption robotaxi service providers could exceed $1.5 trillion in revenue by 2030 with gross profits in excess of $1 trillion.

Let's work through the underlying derivation. Constructive criticism welcomed. The richest income earners spend the most time manually driving, and can command $50 per hour after tax. Higher earners are willing to pay a higher share of after tax wages to win time back. Our research suggests that highest income earners would turn down something less than the equivalent of overtime pay in order to win time back. For other cohorts they buy back time at a discount to what they could otherwise take home. This is a fairly sensitive input to overall market size. That millenials are so obviously willing to trade time for money by hiring doordash drivers rather than schlepping to the takeout counter themselves provides decent anecdata that there is some truth to this curve.

When a consumer decides to take a robotaxi they are not just trading time for money, they are also avoiding the cost of running their own vehicle. Top decile earners spend $.76 per mile, inclusive of the cost of purchasing vehicles, on getting from place to place (excluding air travel). Pretty consistently, by income decile, the marginal cost of mobility runs at ~$.17 per mile. This model assumes that people that already own vehicles are only willing to pay that $.17 at first, plus the value of their time. Over the typical vehicle life-cycle we assume that consumers avoid new vehicle purchases as they grow increasingly reliant on robotaxi. 2 car households become 1 car households and more of the transportation budget shifts into robotaxi. (note that the fixed ownership bumpiness across income decile is almost certainly just an artifact of extracting this from a single year's CEX data crossed with a line item--vehicle purchases--that is infrequent but large across households; I clearly should smooth that but it's not particularly material to conclusions.)

Stack the time value of money by decile, the time spent driving, and the direct spend on driving and you get equilibrium addressable markets by income decile plus the clearing per mile cost. US-wide the clearing cost per mile is $1.36 with highest income deciles willing to pay $3 assuming that they forego vehicle purchasing. If everybody insists on continuing to own a car then the nationwide clearing cost falls to $.97 per mile with highest income earners willing to pay $2.50

A total addressable market of $2.8 trillion that grows to $3.9 trillion as people forego vehicle purchases. Seems largish. But it's always easy to draw a dotted line around an addressable market. What does the likely rollout look like?

Before staging the rollout we have to figure out where the opportunity resides. We adjust zip codes by income per capita and income skew, roll those up into metros and roll those up into states, adjusting metros by average mph (a lower average mph increases consumer willingness-to-pay per hour, since they are saving more manual driving time per mile.) We then score, at a state level, weather friendliness and regulatory friendliness. This provides state level launch stack ranking and within each state we assume that providers will launch the largest TAM metros first. These graphs so dollar per mile (height) by miles available width by state and metro in likely launch order (color-coded by region). You can also get a sense for the wealth contribution to each TAM (where New York is attractive due to low mph, high income per capita and high income skew, but then falls lower in the launch order due to weather and regulatory friction.)

We model adoption on a series of three lagged diffusion curves. First the robotaxi provider has to open up a state. Upon doing so they begin to penetrate each metro area. As each metro area opens up the population begins to take up the technology, sourcing from the highest income earners first, with the effective price in the metro coming down as it penetrates more of the population and more supply comes online.

These rollout curves suggest almost $2 trillion in revenue by 2030, with roughly $1.5 trillion accruing to gross profit using a Wright's Law assumption on the cost to provision service.

Notable that robotaxi doesn't need to roll out to a huge number of metros or states in order to address a substantial portion of the market. Hitting the first 10% of modeled launch metros opens up 40% of addressable gross profit. 25% opens up almost 2/3rds. Even if the long tail proves more difficult to tap (for regulatory reasons or otherwise), there is plenty of capture-able market value at hand.

What does this mean for supply? On 80,000 revenue miles per robotaxi this would suggest the market saturates at 30 million robotaxis, with annual supply additions peaking at 5 million units. Plenty of reasons to believe this is conservative however. This modeling exercise makes no assumptions about flex supply (for peak vs trough) nor any assumption about miles-demand-elasticity. The lack of assumed demand is probably offset by flat-ish pricing even for high end consumers. Net, I would wager that this model overstates the amount of market that lives at a $2.50 per mile pricepoint but understates the miles-driven demand that occurs as prices step down. Higher end consumers will probably be provided the equivalent of a comfort class that they'll take up, but they'll also probably send their kids across town much more often and go to other neighborhoods for dinner or across town for a meeting simply because the effective cost of doing so has been lowered.

At a 15% discount rate, this work suggests that there is $12+ trillion in present value gross profit available in US robotaxi (on these assumptions.) Of course this wouldn't all flow into cashflow. There is a substantial charging infrastructure and servicing buildout that will need to accompany robotaxi scaling, and there will certainly be labor, admistrative and customer acquisition costs that will flow through below the line. Net, it does give a reasonable since of the scale of the opportunity that market participants seem unwilling to confront.

Carving out a conservative case. One could reasonably assert that people won't be so willing to trade off time for money and that there is less upward skew to that data as you move up the income curve. One could also claim that consumer adoption will happen more slowly, that robotaxi suppliers will have a harder time launching the long tail of markets, and that their cost structure will start higher. Under those circumstance perhaps only half the population would be willing to forego car ownership. With those inputs present value gross profit falls to $4 trillion (on a 15% discount rate)

Source / Methods: ARK Invest estimates using BLS ATUS (time use) + CEX (consumer spend) microdata, combined with imputed value of driving time, willingness-to-pay adjustments, and modeled geographic rollout + cost curves. Notes: Illustrative market-sizing model—not a forecast. Includes imputed (non-cash) time value. Results are highly sensitive to assumptions (adoption, timing, costs, discount rate). Disclosure: For informational purposes only, not investment advice. Estimates/assumptions subject to change; no assurance outcomes are realized. Third-party data believed reliable, not guaranteed.

@threadreaderapp unroll