Why Waymo and Cruise Will Go Bankrupt

by

I, Brian Wang, talked to Herbert Ong about the problem with the Waymo and Cruise and robotaxi business model. They both have been losing about $2-3 billion per year for the past five years and were losing $1 billion per year before. Waymo started in 2009 and Cruise started in 2013. Waymo has probably has cumulative losses of $15-20 billion since they started and Cruise has cumulative losses of $10-15 billion.

Robotaxi companies are trying to take market share from Uber and other human driven ridehailing. The robotaxi companies only have a few million in revenue now but the leaders have over $2 billion per year in costs and losses. Cruise, Waymo and the others are paying for thousands of programmers and support staff but not a few hundred drivers for a few hundred robotaxi. They are also buying all of their robotaxi while Uber hires drivers that use their own vehicles.

GM Cruise and robotaxi companies have huge costs for the development, purchase and maintenance of the robotaxi fleet. Uber and human ridehailing incumbants can cut prices in the cities where the robotaxi try to launch. Uber can cut a deal with the third or fourth place robotaxi companies to get their robotaxi. Uber has many advantages and GM Cruise and the other robotaxi companies may never become profitable. GM Cruise and Waymo’s best chance is to get to scale by 2030 by spending $50-100 billion on a large fleet and have it operate perfectly.

If a Robotaxi company wanted to equal Uber revenue today from 6 million cars and drivers, then they would need at least 1 million robotaxi. They are building and owning their own robotaxi. This would cost about $50 billion for 1 million robotaxi. They must have the software development and the support staff. The suppporrt and maintenance staff increases as the robotaxi fleet increases even if the 2000-3000 programmers stays relatively flat. What is an estimate of a staff to maintain a fleet without the drivers? Hertz rents 424,000 cars with a maintenance and sales staff of 25000. Uber only has a staff of 32000 to operate 6 million cars and drivers.

From 2018 to 2021, SoftBank Vision Fund, GM, Honda, Microsoft and others have invested over $8 billion in Cruise. GM’s Cruise robotaxi service has expanded from 70 to 300 robotaxis operating in San Francisco and will soon expand to Phoenix, Los Angeles, Austin and Dubai. GM Cruise had increasing losses of $561 million in the first quarter of 2023. This will be over $2 billion in losses in 2023. GM Cruise will having increasing billions in net losses until they reach profitable scale.

IF GM Cruise grows revenue by 1000 to 2000 times (100,000% to 200,000%) by 2030 and achieves operational and financial efficiency then it would become very profitable. However, they must continue to undercut Uber, taxi and public transit and gain market share while also fighting off Waymo and other robotaxis. This will take perhaps $50-100 billion or much more cumulative losses to finally reach profitability.

McKinsey says OEMs and start-ups have invested $106 billion in autonomous-driving capabilities from 2010-2021. Much of this funding has gone toward enhancing advanced driver-assistance systems (ADASs), which handle braking, object detection, and other critical vehicle functions.

GM Cruise will have an annual net loss of $2 to 3 billion in 2023 and will likely surpass the ten billion per year losses that Uber had in 2021 before Uber crossed over into profitability.

Here we will go over the current Cruise business and business plan and compare it to Uber and look at GM Cruise plans to about 2030.

Revenue from 300 robotaxis is about $5 to 20 million per year. Cruise is driving about 1 to 2 million miles per quarter. Reports have indicated that Cruise robotaxis in San Francisco currently have less than 10% utilization. This means only 10% of miles are paid. Eventually, this could rise to near 50% utilization.

Cruise charges a base fee of $5 and a $0.90 per mile and $0.40 per-minute rate. Fares also include a 1.5% city tax and are calculated using the estimated time and distance of the fastest, most optimal route. This is about $12 for a 2.6 mile trip and $18 for a 5.2 mile trip. The San Francisco Transportation Network Company (SFCTA) reports the average robotaxi trip in San Francisco is 2.6 miles. The trips should get longer on average as the area of operation expands.

Cruise is currently charging customers about 20% lower prices than Uber for the same trips.

Ridehail vehicles should be able to eventually average around 26 to 29 trips per day in San Francisco and perhaps other cities. The number of trips for Cruise is still much lower. They could reach between 16 trips per day with the expanded service area and expanded fleet for the second half of 2023. 25 paid trips at 2.6 miles per trip per day would be 100 paid miles per day. This would be $300 per day per robotaxi. This would be about $100k per vehicle per year in revenue.

GM Cruise plans to reach $1 billion per year in revenue in 2025. This would be 100 times more than the potential revenue runrate in Q4 2023. This would need about a global GM Cruise fleet of 10,000 robotaxis.

If Cruise gets to 10000 self driving cars operating for in 2025-2027 in about ten cities making 25 trips per day then they will make
10000 robotaxi
* 20 trips per day
* 90 days in quarter
* $14 per trip.
$252 million per quarter.

This would be 200000 trips per day versus about 2500 trips per day in first half of the year (2023). This would still be over 100 times fewer trips than Uber did in Q2 2023.

GM Cruise costs would increase. 9700 more robotaxis would cost about $1 billion. In 2023, Cruise has about 2000 employees developing and supporting the software, hardware and operating the vehicles. They do not have permission to operate without a safety driver in many locations and they need people to move robotaxis that get stuck or have problems.

Uber has a staff of 32000 in 2022-23. Six million Uber drivers are not Uber staff.

Uber Versus GM Cruise Quarterly Miles and Revenue

In Q2 2023, Uber booked $33 billion in a quarter and the drivers and couriers make about $15 billion and the company makes $9 billion and partners make $9 billion. This about 10 billion trips per year, 2.3 billion trips in the quarter or about 25 million trips per day.

In theory, GM Cruise booking revenue at half of Uber would mean the same profits. IF GM Cruise had tiny flat robotaxi software costs. GM Cruise would not pay drivers and theoretically can keep operational support staff at a fraction of the cost of drivers. GM Cruise at $1.5 billion per year in revenue would be equal to Uber in 2014.

GM Cruise might crossover to profitability in the range of $50 to 100 billion per year in revenue. However, Uber could get 20-30% more efficient or profitable and the new robotaxi’s never beat Uber. GM Cruise plans to reach $50 billion per year in revenue in 2030. This would require 500,000 robotaxis and 4 million Lidar units. Global lidar production would need to increase about one hundred times. The robotaxi fleet would cost $25-50 billion for the vehicles. The annual operating losses would increase from the $2.5 billion in 2023 to $10-20 billion per year in 2029.

In order for Cruise and the other robotaxi companies to take market share at scale from Uber and Lyft then they will need to offer both a better service and a discount on costs. They will need to offer more than the current 20% discount versus Uber pricing. Cruise will need to operate vehicles for more miles, more paid miles to reduce costs per mile and have the support staff become far lower than the displaced cost per unneeded human driver. Supporting strong growth will likely mean undercutting Uber pricing by 30-50%.

Until Cruise and other robotaxi reach significant scale, they will make less than Uber on in-vehicle advertising.

GM Cruise, Waymo and Apollo will lose more than the $2 billion per year in losses as they scale to the $25-70 billion per year in revenue needed to achieve profitability. Like Uber, they will lose money for years until they reach profitability. The hope and plan is to reach this around 2030.

Uber Gross Bookings grew 16% year-over-year (“YoY”) to $33.6 billion, or 18% on a constant currency basis. Uber Mobility Gross Bookings of $16.7 billion (+25% YoY or +28% YoY constant currency) and Uber Eats and other delivery Gross Bookings of $15.6 billion (+12% YoY or +14% YoY constant currency).

Uber Trips during the quarter grew 22% YoY to 2.3 billion, or approximately 25 million trips per day on average.

Uber company Revenue grew 14% YoY to $9.2 billion, or 17% on a constant currency basis. Uber Drivers and couriers earned an aggregate $15.1 billion (including tips) during the second quarter of 2023, with earnings up 17% YoY, or 19% on a constant currency basis.

Uber Income from operations was $326 million, up $1.0 billion YoY and $588 million quarter-over-quarter (“QoQ”).

Net income attributable to Uber Technologies, Inc. was $394 million, which includes a $386 million benefit (pre-tax) primarily due to net unrealized gains related to the revaluation of Uber’s equity investments.

In March, 2023, Cruise was forced to recall 300 Cruise AV units following a crash with a city bus in San Francisco.

GM is targeting $1 billion in revenues by 2025 as it plans to ramp up expansion plans in 2023. However, this will not mean profitability in 2025 when Cruise reaches $1 billion per year in revenue. Cruise will at least to have to reach $25 billion per year in revenue to reach profitability. This is especially true if Cruise follows the plan of rapid expansion and spending tens of billions each year on expansion. Cruise had about 300 autonomous vehicles at the end of 2022 and in mid-2023 they have about 400 now with about 3000 operating a paid commercial service in San Francisco.

Software and programmer developer costs will increase form $2 billion per year to $5 billion per year or more. The robotaxi software will have to work flawlessly in over two hundred cities.

The various scaling challenges:
The software will have to work better in hundreds of cities.
Any operational issues and costs like maintaining hyperaccurate and real time updated maps will have to be maintained.
The larger robotaxi fleet will have to be maintained.
The supply of Lidar and other sensors will have to be secured and the supply does not yet exist for the planned scale.
Market share will have to be taken from existing providers and will have to taken from other robotaxi competitors.
There will only be two main robotaxi companies for each region. One for North America and one for China and perhaps one for Europe and another for the rest of the world.

The robotaxi system will have to work for all of the desired cities and become profitable and dominant market share is needed. All supply chain and regulatory hurdles must be overcome.

Tesla FSD and advanced driver assist has a different business model that is already profitable. Tesla’s FSD business model and future robotaxi plans can further reduce market and profits for pure robotaxi companies.

7 thoughts on “Why Waymo and Cruise Will Go Bankrupt”

  2. Elon said FSD will become aware ,this Fall,and he’s never been wrong. I guess GM,doesn’t believe him.

  3. Will it be possible for a remote driver to safely take control of the car when it gets stuck in a situation that a human driver could easily get out of? This is the only way I see for Cruise or Waymo to overcome the foreseeable future of unpredictable and imperfect software freezing up at what appears to humans, as random incidents. Sending a human driver out to scramble into the front seat to drive a few feet whenever the AI decides to stop in the middle of the street is neither safe nor sustainable when passengers are rightfully annoyed or scared with moving traffic dangerously swerving to avoid their stopped car.
    OTOH, a few seconds of inconvenience while a remote driver takes over a stopped car to get it moving safely again – probably with a full fare refund – might be acceptable, for a while anyway. If it doesn’t get a lot more reliable in a few years, GM etc. may run out of money for this experiment.

  4. Cruise and Waymo robotaxis should become profitable before 2030, and in so doing. eventually be hugely beneficial for everyone. Because robotaxis will advantage everyone, government must not let Cruise or Waymo fail.

  5. the current offering is a bit ahead of the tech and the market. 1. I suspect there will soon be a dedicated EV platform for robotaxis that will seriously cut production and operation costs. Capital expenditure will decrease greatly soon. this might also lead to service levels for customers requiring different distance and interior amenities. single and double occupant cars could be called instead of the large 4 passenger cars shown. 2. FSD and other tech advancements will make trips safer and more efficient. 3. as the population transitions, owning a vehicle will become less and less relevant. there is a driving culture in advanced countries that will fade and become based more on requirements than status. amazon was always going bankrupt and no one was going to buy online 25 yrs ago. and now?? these companies might stumble, but go bankrupt is too early to say.

  6. A robotaxi will never rape its passenger. Women leaving the bar will choose the robotaxi every time over a human driver. A couple of Uber rides ago my driver decided to tell me and my buddy about his ideas on women. I didn’t realize they were all aware of their sexual power over men and conspired to use it to humiliate and control the male species. I bet my future ride in a robotaxi won’t include misogynistic incell conspiracy theories coming from the empty drivers seat. In the future people will be able to rent out their robocars and robotrucks when they are in the office, sleeping, or whenever else they are not using them. There will be a multitude of apps to directly link people who need a ride to folks with idle smartcars. Waymo and Cruise will get to profitability first. Their brand recognition will count for something for a while, but with the advances in AI/ML continuing to grow, in 10 years the tech will be cheap enough that anyone will be able to create their own robotaxi service. Comma.ai already sells a single camera driver assist system that works on 250 vehicles for $1250. Their code is on Github and its open source. It won’t take long for projects like theirs to become integrated with Machine Learning and other new parsimonious 3d imaging advances like NeRFs to create a lot of competition in the self driving vehicle economy. What a fantastically thorough article Brian, I really enjoyed it.

  7. Cruise and Waymo will keep burning Capital competing with Uber and Lyft until Tesla FSD is ready to pass regulatory checks – then Tesla Network will flood the global market with millions of superior robotaxis (millions in new production + millions already on the road) that don’t need hidef mapping or Lidar and put all of them out of business pretty quickly.

Comments are closed.