Right-Sizing Robotaxi Fleets, Urban Freight, and Two Collapses: Antitrust and Tesla's Brand
Off-Ramps for 3 April 2025
Welcome to Off-Ramps! Today I’ll highlight four interesting pieces that I think you will enjoy reading. Please enjoy these on your morning commute, or save them for your weekend.
Right-Sizing Robotaxi Fleets
Recently, Bill Gurley, venture capitalist and former Uber board member, asked a very important question.
He put that question so briefly that it may not be obvious why it is important. Let’s unpack it, and then see if we can find an answer.
In any given city, demand for rides is not uniform throughout the day. There will be a certain base demand, but it will predictably spike during the morning and evening commutes. At night the base will fall, but will spike again as bars begin to close, or for special events, like immediately after a Taylor Swift concert ends.
This lack of uniform demand is a problem, because it means no matter what a robotaxi firm (in this example, Waymo) does, it will suffer. If it right-sizes its fleet to peak load, it will have too many vehicles during the off-peak. That’s problematic, because robotaxis will be very expensive, and this model guarantees that much of the time, some of those robotaxis will be stranded assets, earning no revenue to offset their cost. Another solution would be to lower the price of rides, to elicit more off-peak demand; but doing so would erode its value proposition, while also contributing to off-peak congestion, which will anger its host cities.
Conversely, if it builds its fleet to meet base demand only, peak demand will go unserved, leaving money on the table while angering its customers, who won’t be able to use the service.
Historically, ridehail solved this problem by not owning its own fleets, and putting that risk onto the drivers who made their services available on the platform. When supply of drivers exceeded demand, that was the driver’s problem, not Uber’s; Uber didn’t own the vehicles, so it incurred no costs for the driver’s time nor the wear-and-tear on their cars. And when supply of drivers fell below demand, Uber could implement surge pricing to encourage more to get on the road.
In an interesting essay, Reilly Brennan argues that Waymo can’t take advantage of this situation, but Tesla can. The Cybercab model, where private owners release their vehicles to the system when they choose, provides a pool of supply that can be drawn upon, perhaps via a surge mechanism, when demand requires it. He concludes, with some grandiosity, that “how [Tesla] handles a supplemental supply of individually-owned vehicles is the most impactful [moment] for the global ridehailing market. Whoever can successfully build and manage a robot fleet with dynamic ownership can take over the world.”
As I have argued before, and often, Tesla’s plan for private owners to use their vehicles for double duty, as personal vehicles sometimes and as robotaxis at others, is unlikely to succeed. Consequently, I don’t share Brennan’s certainty that Tesla has a silver bullet here. Instead, I put my expectation in a different scenario he offers but doesn’t explore: namely, that even as Waymo soaks up base demand, there will still be demand during the peaks for privately-owned, human-operated ridehail vehicles. It is the most obvious solution to Gurley’s riddle, and it aligns neatly with the argument my co-authors and I make in The End of Driving: that even in a world of ubiquitous robotaxi, some things will change less than we might expect.
Even in that world, there will be some demand for human ridehail drivers.
Preferences in Urban Freight
In my recent interview of Sandra Rothbard, we explored the ways that the pressure of next-day or even same-day delivery is applying stress to urban infrastructure that was not designed to accommodate such pressures. The result is double-parked trucks, blockaded bike lanes, overfull parcel rooms, and air pollution.
Sandra proposed a variety of solutions that cities should implement: micro-hub consolidation centres, infrastructure optimized for freight—so smarter loading docks and wide bike lanes—and dedicated city staff to oversee it all. But one suggestion we didn’t cover is discussed in a recent report from McKinsey (of all places): discounts for slower delivery.
McKinsey, notoriously, is in the business of telling its customers what they want to hear, so their recommendations should always be treated cautiously. But in this case their analysis isn’t obviously aimed to flatter anyone, and as such is worth considering. As per the firm’s report, U.S. consumers are falling out of love with speedy delivery in favour of cheap delivery, even if that takes extra time.
(Extra time, in this context, means that city-dwellers may be willing to wait as long as three days for their packages; a far cry from the standard ‘allow four-to-six weeks for delivery’ of my youth.)
The reason for this change in preference isn’t clear, but I imagine stems from both the heightened inflation of the past two years, and increased anxiety about economic turbulence to come. There are very few goods where a wait of an extra day or two imposes severe hardship, and if waiting that time permits a discount, cost-conscious consumers are willing to incur that time penalty.
This gives me hope, because many of the problems that Sandra cites with urban delivery are functions of time being the key variable. Even an extra day of time would allow logistics firms to consolidate deliveries, plan more efficiently, and take advantage of amenities that save money; urban consolidation hubs being only the most obvious.
So I hope what McKinsey is reporting here is correct. A world where only the most urgent goods are shipped same-or-next day is a world where more people get better outcomes, and I hope consumer preferences are shifting to make that vision come true.
The Deflation of Antitrust
As I’ve discussed before, abundance is the Hot New Thing. Klein and Thompson’s book of that name was the media star of March 2025. I used it as a way to think about the ongoing Canadian election. In that vein, it’s being talked up as a new organizing principle for the Democratic Party, now that identity politics seems to have lost its power to persuade.
In response, the partisans of the Hot Old Thing are pushing back hard. That Hot Old Thing, in a word, was antitrust.
By that, I don’t mean solely its textbook meaning, namely opposition to monopolist rentiers, who use economic power to impoverish everyone for their own benefit. I mean its larger sense: opposition to corporate power generally. On that view, associated somewhat with Bernie Sanders and especially with Elizabeth Warren, the principal problem in American life is that large corporations not only have economic power, but also political power, and as such are able to impoverish others, and make them less free, in order to aggrandize themselves.
This wing of the Democratic Party had significant influence on the Biden Administration, and looks to have influence again in a post-Trump world. They are not prepared to give up pride of place to ‘abundance’.
Unfortunately for them, they will have to do so. Noah Smith lays out a convincing argument that progressives need a new narrative because the old one, which places blame for everything on corporate power, won’t do. The anti-corporate agenda was unpopular, for one thing: the very people it was supposed to benefit ignored Biden’s efforts in this regard and defected to Trump. Nor will it solve the problems of the day: it remains the case that red states like Texas, with their presumably-corporatist alignments, have more housing growth and more sustainable energy than blue states like California, which lacks such an alignment. This suggests that the reason America can’t build isn’t to do with corporatism, but instead with, as Klein and Thompson have suggested, with regulatory sediment that chokes abundance.
Progressives should recognize this, and act accordingly.
How Tesla Might Die
It seems like every time I do one of these link round-ups, I include something about the slow-motion collapse of Tesla Motors.
I’m not happy about this. I drive a Tesla, and I think it’s a fine car. I have brand loyalty.
But I also have eyes to see and ears to hear, which keep telling me this is a brand in serious trouble. I’ve talked before about how its international reputation is tanking, and Chinese EVs are outcompeting it overseas. But now the chickens are coming home to roost in the USA itself.
Tesla showrooms are the target of organized protests, as are Tesla Supercharger stations and even Tesla vehicles owned by private citizens. Democratic state governments like New York are ordering hostile audits. Technocratic centrists like Matt Yglesias are now urging the Democrats to reorganize their policy commitments so as to deal Tesla out and competitors like Waymo in. Meanwhile, the surge in market capitalization that Tesla enjoyed thanks to its association with the Trump administration is over (though it is enjoying a slight rebound, presumably on the market’s assumption that it is the automaker which will be harmed the least by the new tariff regime).
And now two-thirds of Americans say that they will not consider buying a Tesla car.
What does this mean? Will Lockett thinks it means that Tesla is going to die.
Lockett rehearses many of the arguments I’ve made above: the firm’s sales are down and it is unpopular. But he adds to this a look at Tesla’s financials, which are damning. Its valuation implies a profits/earnings ratio of 68, which is ludicrously high compared to the next-most-valuable automaker, Toyota, which has a P/E ratio of only 8. Stipulate that Tesla is overvalued, and one should be extremely concerned, given the firm’s $48 billion in direct debt, and its hidden debt (Elon Musk has used substantial amounts of his Tesla shares as collateral on other ventures). This leveraging makes Tesla extremely vulnerable to downturns; if the stock value starts to fall, it could trigger margin calls that can’t be met, leading to the firm entering a death spiral.
Lockett’s headline says this is how Tesla will die. I’m not so sure; there are ways out. Trump has already shown a willingness to intervene; he might again. The board might make drastic moves. Nothing is certain.
But the idea that the firm will continue be the growth company of investor’s dreams, or that Elon Musk will, courtesy of his Tesla stock, continue to be the world’s richest man, is uncertain as well. Adjust your expectations accordingly.
>>Household vehicles were driven an average of 64.6 minutes on a typical day in 2022 (including all trips made that day) and parked for the remainder of the time (95%).<<
https://www.energy.gov/eere/vehicles/articles/fotw-1356-august-19-2024-household-vehicles-were-parked-95-typical-day-2022
It seems clear to me that Waymo should prepare for peak demand. Driverless cars are fundamentally different from cars with drivers, because there is no driver to pay. Owners of private automobiles are not losing sleep because their cars are parked 95% of the day. And they're not worried about the interest on the money they paid for the car which continues even when the car is not in use. The decision to purchase a car and bear that interest has already been made because of the utility of having a car available when thry want it. After that decision has been made, the only cost that owners of private vehicles are concerned about is maintenance. And vehicles that are not moving are not degrading.
So if private owners are quite happy with a situation where their vehicles are not being used 23 hours of the day, that Waymo should be happy with a situation where their vehicles are used only half of the day.
Arguably, in the long run, services like Waymo will not be competing with Uber and Lyft, they will be competing with private vehicle ownership. I was curious, so I asked OpenAI Deep Research for a report on usage levels of passenger vehicles overall (https://chatgpt.com/share/67eea043-9be4-800b-9a9a-2ef70ce9ef7e). The results indicate that passenger vehicles in the US are overprovisioned by roughly 10x vs. peak usage. Thus, if Waymo were provisioned for 1x peak usage, it would get roughly 10x better utilization than today's overall vehicle fleet.
I think I may be saying more or less the same thing as Ethan's sibling comment, just coming at it from a different direction – the fact that the status quo of private vehicle ownership is economically viable implies that the need to serve peak usage periods should not push costs out of bounds.
Of course, I'm glossing over all sorts of details here, but in the big picture it seems like it should be economically feasible to provision a Waymo-like service for peak usage, once the cost of the AV hardware comes down a bit? Though I could imagine that in the current early-adoption period the peak-to-mean ratio for Waymo is higher than for passenger vehicles overall (e.g. relatively high usage during events such as a big concert, relatively low usage during routine commute hours).
(Caveat, I made zero attempt to investigate the figures Deep Research produced, so they could be way off.)