Uber unleashed: 12 times the money, 20 times the cars
At last week's quarterly earnings Uber outlined projections for the proliferation of autonomous vehicles on our streets that is, frankly, frightening. In their 'Autonomous Vehicles Spotlight' briefing, Uber estimates, when finally unshackled from pesky workers, US market revenues from autonomous vehicle services will grow to USD $1 trillion, a 12 fold expansion from today's position. And with Uber predicting a collapse of pricing from current fares of around USD $2.50 per mile today to just USD $0.50 per mile for fully autonomous vehicles, this will lead to a massive expansion of vehicles for private hire on our streets.
To illustrate, if we use Uber's reported trips and revenues for 2024 which is around 50% of the global total, and assume each vehicle carried out 6 trips per day on average, we can estimate that Uber's US fleet size in 2024 was around 2.6 million vehicles active daily on the platform. And while a fully autonomous vehicle, minus the worker, can work for longer hours, there are inevitable constraints. As Uber mentions in the statement:
AVs need to be charged multiple times a day and serviced monthly. AVs will also require consistent cleaning and available parking.
So if we take a conservative approach and estimate that an average autonomous vehicle might carry out 18 rideshare trips per day compared to say 6 trips per day today, that would mean a daily requirement of 52 million vehicles on the road daily in the US, a 20 fold expansion of the fleet. And with fares dropping by 80% to $0.50 per mile, the average trip fare would drop from $14.44 reported in 2024 to just $2.89 when the autonomous vehicle rideshare and delivery services market matures.
Uber's $1 trillion vision for driverless cars in the US
2024 | At AV market maturity | Growth | |
Estimated Uber US Gross Bookings* | $81,386,500,000 | $1,000,000,000,000 | +1129% |
Estimated Uber US Trips* | 5,636,500,000 | 346,279,788,417 | +6044% |
Estimated Uber average US daily vehicle requirement** | 2,573,744 | 52,706,208 | +1948% |
Uber US price per trip | $14.44 | $2.89 | -80% |
Source: Uber Technologies, Inc. Q4 2024 Earnings Autonomous Vehicles Spotlight
*assume 50% of global total in 2024, source: Uber Q4 2024 financials, Uber US & Canada revenues were 55% of total when Uber last reported regional revenues in Q2, 2023.
**assume average of 6 trips per day per vehicle in 2024 and 18 trips per day per vehicle at AV market maturity
London locked: 4 million autonomous vehicles
If we apply Uber's estimates for the US to the UK market, we would also expect to see Uber fares collapse here and the fleet expand exponentially. Uber's anticipated 80% drop in fares would see a current fare of around £32 from Paddington to Heathrow drop to just £6.40.
If we apply the same growth projections here then the current London private hire fleet might expand from around 100,000 vehicles today to around 2.1 million vehicles as the autonomous vehicle service market matures.
Exacerbating the problem is that autonomous vehicles will, on average, spend more time on the road per day than manual drivers constrained by the limits of human endurance. So while we could see 1948% expansion in the number of vehicles, the number of trips taken in what is already Europe's most congested city would expand by 6044%.
And whilst London will be transformed by a twenty fold expansion of the minicab fleet, this would only be half the story as a potential expansion of the autonomous delivery fleets of Uber Eats, Deliveroo and Just Eat gets underway at a similar scale and pace.
Uber's driverless dream for London
2024 | At AV market maturity | Growth | |
Uber UK price per mile* | £2.00 | £0.40 | -80% |
London TfL licensed fleet** | 97,000 | 1,986, 366 | +1948% |
London estimated app delivery fleet | 100,000 | 2,047,800 | +1948% |
Total London on-demand vehicles | 197,000 | 4,034,166 | +1948% |
*Based on current Uber fare quote of £32 for Uber X service from Paddington to LHR.
**Source: TfL
Waiting for Godot - worker exploitation continues
In our recent Dying for Data report we exposed that Uber's current utilisation rate is as low as 40% in London compared to 55% utilisation in New York where, unlike London, capacity is controlled by regulation. Contrary to multiple court rulings, Uber continues to refuse to pay workers for waiting time which industry wide amounts to £1.9 billion in wage theft from workers.
But the actual situation today is perhaps even worse than we conservatively calculated when we produced the Dying for Data report. In Uber's spotlight communication to investors they finally admit that..
..in a typical large city, a fixed fleet designed to meet the weekly peak will have up to 95% of vehicles idle during the multiple weekly troughs.
But because Uber unlawfully refuses to pay workers for their standby time, drivers spend many hours unpaid just to ensure Uber has a reliable network service.
Conversely, a fleet sized below peak cannot deliver the reliable 4-minute ETAs that consumers expect. This dilemma is compounded by the fact that vehicle supply will need to vary dramatically throughout the year: in many North American cities, demand peaks around March and trends down through the summer, only rebounding in the fall. Weekly peaks may be 50%+ higher than the lowest week of the year. This could mean that a fixed network that is well supplied in March may find itself idling a third or more of its vehicles for several months, until Halloween. Because demand patterns look similar (although not identical) between cities, underutilized vehicles cannot be easily repositioned and will likely sit unused for months, generating additional cost and complexity—against zero revenue.
Uber drivers have been mercilessly exploited for years and are forced to eat the cost of under utilisation while Uber over supplies their network for the strategic dual advantage of rapid customer response while driving down wages. The under utilisation issue is only now a concern because Uber is promising to help future fleet investors optimise so they can get a return on their assets that workers that workers were denied for the labour.
Any standalone player with a fixed depreciating asset will need to build against that reality: choosing between running a highly underutilized network (if supply is built for peak demand) or a highly unreliable network at peak periods (if supply is built for anything less than peak).
It is a shocking betrayal that only now is Uber coming clean about its dreadfully wasteful and exploitative business model that cheats workers out of standby pay. It is insulting that the admission is only now made as a byproduct of a warning to future investors about the hidden costs of under utilisation. It is a cost now carried by low paid workers but in future would have to be borne by institutional investors and Uber is prepared to help them optimise to maximise payment - something they fail to do for workers. Uber has all respect for capital and none at all for labour.
Surely the regulators will save us, won't they?
But surely regulators in the UK would halt the swamping of our streets with autonomous vehicles? Unfortunately, the signs for restraint are not good. Sadiq Khan last year reversed his long standing support for a cap on private hire vehicle licensing.
At any rate, section 83 of the newly minted Automated Vehicles Act 2024 makes provision for the 'disapplication' of existing taxi and private hire regulation. FOI's show that the briefing notes prepared for a January 2024 meeting between Transport Minister Stephen Harper and Uber's General Counsel Tony West offer Uber assurances of a more 'flexible' approach to regulation of autonomous vehicles.
The industry titans are already lobbying hard to make sure there will never be a cap on autonomous vehicles in our streets. In 2022, Waymo told the Law Commission inquiry into regulation of autonomous vehicles that:
...we believe a national scheme should allow for broad commercial deployment. Creating limits on numbers would impede this and would be akin to imposing a second pre-deployment trial phase. It would also create an added layer for decision-makers to govern and industry to comply with.
In a 2019 paper now removed from it's website TfL warned of a dystopian London cityscape of the future with 'ghost' vehicles driving around empty in algorithmic search of work. Certainly, there is already a significant risk of community backlash against such driverless operations as we have seen in San Francisco recently. So in 2025 is TfL now ready for the challenge? The signs are not good. This year, TfL published a new document called 'TfL’s approach to automated vehicles' which is, frankly, thread bare on policy solutions to the clear challenges ahead.
The interests of Britain Plc are further tilted in Uber's favour with their substantial if undisclosed equity investment in Wayve, a Cambridge based tech company specializing in software development for autonomous vehicles. As the current UK government forges ahead with a let-it-rip approach to AI, its hard to see where the guard rails are to prevent the excesses of a wild west tech sector gold rush to take over our streets.
It is unfortunate that the UK government thinks that less regulation is permissible, not least because autonomous vehicles present a significant threat to personal and community privacy. Vehicles roaming the streets 24/7 hoovering up community location data are an intrusion but the invasion of personal privacy of partners could be extreme. For example, a 2018 patent filed by Uber, though since abandoned, is directionally instructive. The patent filing anticipated the collection of biometric data from autonomous vehicle passengers to determine their psychological and physiological state, again, something that is profoundly unlawful to do in Europe according to the GDPR.
.......can include heart rate (e.g., heart beats per minute), blood pressure (e.g., systolic and diastolic pressure), grip strength, blink rate (e.g., blinks per minute), facial expression (e.g., a facial expression determined by a facial recognition system), pupillary response, skin temperature, galvanic skin response, gestures (e.g., head, torso, hand, arm, foot, and/or leg gestures), amplitude of vocalization (e.g., loudness of vocalizations/speech by one or more passengers), frequency of vocalization (e.g., a number of vocalizations/speech per minute), and/or tone of vocalization (e.g., tone of vocalization/speech determined by a tone recognition system).
So what is Uber's autonomous game anyway?
Dara Khosrowshahi's decision to sell off Uber's own autonomous vehicle unit in 2017 may yet be terminal to his career at Uber. In 2024, the chickens came home to roost and Uber was put under pressure to answer the threat posed by Tesla and Waymo.
So, what is the strategy? First, Uber wants to position itself as the market place leader where all future fleet owners can provide vehicle capacity for sale. Uber will source demand, collect payment, charge & clean vehicles and do customer support.
Second, Uber is pursuing what it calls an asset-light strategy for autonomous vehicle development through key partnerships with Wayve and Nvidia. Tesla and Google are rather more chained to heavier tech reliance on video, lidar, radar and rules whereas Wayve appears to be more nimble in also using sensors but relying more on real time AI and deep learning from sensed and historical data. Crucially, Wayve's tech is manufacturer agnostic for now and, in taking an AI first approach, the hope is it will prove to be more cost competitive in the long run. Uber, Nvidia and Wayve are tightlipped about exactly what they are working on but it is almost certain that Uber is opening up its vast store of data (likely unlawfully if UK & EU driver data is being transferred for this purpose) to help train Wayve software models superpowered by NVIDIA compute.
In the 2010's Uber outfoxed auto manufacturers who thought they might control the emergent rideshare market. In the 2020's Uber is out to do it again by partnering with Wayve and NVIDIA. They've already seen off GM's Cruise for AV just as they saw off Daimler & BMW's FreeNow for rideshare. Waymo and Tesla may have other ideas.