Didi Makes $1 Billion Profit While Uber Bleeds — Good News For Robotaxis

Didi Chuxing, the ride-hail service which forced Uber

UBER


UBER


UBER


UBER


UBER
out of the Chinese market (giving Uber equity) has, according to sources, earned a profit of over $1B on their ride-hail business in 2020. This is remarkable both because of the way Covid crushed most ride-hail companies in 2020, and also because Uber — which did face that crush — continues to bleed money. While it reports for Q4 on Feb 10, the Q3 report showed a $1.1B loss for just the quarter.

While one can make many comparisons between Uber and Didi and their styles of business and markets, the interesting thing about this news is how it demonstrates the capability of a ride-hail company for profits, even in difficult times. (Didi managed to grow their ride business in 2020, while Uber and Lyft

LYFT


LYFT


LYFT


LYFT
only survived by growing their food delivery services.) This, in turn, informs about the potential for profitability in the robotaxi business, which many players have bet many billions on.

Uber’s inability to make money has caused some to suggest it’s in a bad business. Uber maintains it only loses money because it spends greatly on growth. In particular it spends a great deal on driver subsidies to recruit and maintain drivers as it grows.

A typical Uber ride works out to about $2/mile, though the formula is more complex. Uber takes 50 cents of that and gives $1.50 to the driver. The driver bears 30-50 cents/mile in car costs and depreciation, and pockets the rest for her time. Uber doesn’t really do a lot for their 50 cents — they insure the ride, and manipulate some bits in their servers. While they have lots of non-recurring costs, the gross margin on that 50 cent cut should be quite high — and yet the profits have failed to appear.

Driving your own car also costs about 50 cents/mile when you factor in depreciation and all other costs. That’s a lot less than the $2/mile to hire an Uber. Many predict that the incremental cost of running a robotaxi should not be too much different than that 50 cents, and in fact could be quite a bit less, both because of profit margins in the 50 cents, and the ability to make low cost single-person robotaxis that cost much less to run but can handle 80% of all trips, which are solo.

But if Uber can’t make money charging you 50 cents/mile with minimal “cost of goods sold” then how does a Robotaxi make money at that price if it also has to pay 20-40 cents for the COGS of the ride? Robotaxis want to be able to be seen as an alternative to owning a car, and as such they don’t want to cost twice as much. They can cost a little more since you don’t have to drive, park or maintain them, but there is a limit.

Didi operates in China, and it’s cheaper than Uber in the USA. In Shanghai, expect to pay about $2.30 for flag drop and 84 cents/mile. Uber varies around the USA, but in Silicon Valley, it’s a similar $2.20 flag drop plus $1.60/mile and a per-minute fee and ride fee. Costs are lower in China, but Didi is making a nice profit with these lower prices.

Like UberX, Didi uses cars owned by drivers. Those cars are cheaper in China, to be sure, as are several other factors. One of the big savings is the drivers often don’t look at the total cost of using their own car, but only the incremental cost. Driving extra costs maintenance and depreciation, but drivers don’t see it. They mostly see costs like fuel and feel they’ve already paid for the car. This is partly true — there are some elements of a car that age by the year, but most age by the mile. This helps ride-hail companies make drivers believe they are earning more than they are. A commercially owned fleet, like a robotaxi fleet, does not have that luxury; it must cover all costs.

Didi’s typical commission is 19% — Uber now charges 25% but earlier drivers can pay less. This means Didi’s profitable service pays about 68 cents/mile for a car and a driver. This suggests Robotaxi profitability is a reasonable prediction, though doesn’t yet give proof about the eventual goal of being price competitive with owning your own car.

The good news is that it’s quite possible a robotaxi will cost less than a regular car. While it may need a few thousand dollars of special sensors and computing, it takes out the huge fraction of a car’s cost that is there for the driver — wheels, pedals, dashboards, fancy adjustable seats, mirrors and much more, all of which cost more than the target price for the sensors. At first, the biggest barrier will be the billions of investment required to bring the vehicle to market, but investors seem happy to provide that.