On 21 October 2026, hours after Tesla’s third-quarter results came out, Ark Invest’s Tasha Keeney told Bloomberg that Tesla’s future is the robotaxi, not the car it sells you. Her boldest claim is about cost: she argues the Cybercab will be cheaper to run than the next-generation vehicle from Waymo, which leads the market.
Key takeaways
- Ark’s Tasha Keeney says Tesla is better placed than its rivals and that the Cybercab beats Waymo’s sixth-generation robotaxi on running cost.
- Ark’s own research puts the Cybercab’s incremental cost per mile at roughly 50% below Waymo’s next-generation vehicle.
- Tesla’s Q3 2026 adjusted earnings of $0.48 a share beat the $0.44 estimate, but revenue of $28.05 billion came in just under forecasts.
- Waymo is far bigger today, and Tesla’s Cybercab faces a federal safety audit with answers due on 30 October 2026.
What did Ark’s Tasha Keeney actually say?
Keeney is Ark’s director of investment analysis and leads its research on autonomous technology and robotics. Her message on Bloomberg was simple: robotaxis are where Tesla’s value lies, Tesla is ahead of its peers, and the Cybercab costs less to operate than Waymo’s next-generation vehicle.
None of this is a new position for Ark. In an April 2026 essay, Robotaxis Are Coming For Your Commute, Keeney estimated that human-driven ride-hailing in the US costs about $2.80 per mile, and that robotaxis could get down to about $0.25 per mile at scale. Ark sees about $34 trillion in enterprise value (the total worth of the companies involved) building up around autonomous platforms by 2030.
Ark says itself that these forecasts reflect its own positive bias and may not prove accurate. Tesla has long been one of the firm’s best-known holdings, and you should keep that in mind.
Did Tesla’s Q3 numbers support the robotaxi story?
In part. According to Yahoo Finance, Tesla reported adjusted earnings per share of $0.48 for the third quarter of 2026, ahead of the $0.44 estimate. Revenue was $28.05 billion, just below the Bloomberg consensus of $28.13 billion.
The surprise was cash. Tesla generated $1.53 billion in free cash flow, while analysts had expected it to burn cash, largely because capital spending of $4.73 billion came in far below forecasts. Even so, Tesla still expects to spend more than $25 billion in 2026 on Optimus robots, AI computing power and the Cybercab ramp-up.
The car business itself is still soft. Deliveries were 486,532, down 2% from a year earlier, and profit fell sharply, as we covered in Tesla Profit Falls 43% as Revenue Stalls at $28 Billion. On the call, Musk said the Austin Cybercab fleet “has grown by a factor of 10” and called full self-driving “the biggest driver of demand”.
Is the Cybercab really cheaper to run than Waymo?
On paper, it could be. Ark’s research suggests the Cybercab’s incremental cost per mile (the cost of driving one more mile) is about half that of Waymo’s sixth-generation robotaxi. That vehicle, called Ojai, is built by Zeekr in China and fitted with Waymo’s sensors in Arizona.
The argument comes down to design. The Cybercab is a two-seat car built from scratch to be a taxi, with cameras only and no steering wheel or pedals. Waymo relies on lidar (laser sensors that map the car’s surroundings) and a costlier sensor suite. Waymo has cut that suite to 13 cameras and four lidars on Ojai, but it is still more hardware than Tesla uses. Think of a food truck built for the job next to a family van with a kitchen squeezed into the back.
There’s a catch. Cost per mile only turns into profit if the cars stay busy and the software is trusted to drive without a human backup. Cheap miles do you no good if a regulator parks the fleet.
Tesla robotaxi vs Waymo: how the two compare
| Factor | Tesla Cybercab | Waymo (Ojai, sixth generation) |
|---|---|---|
| Sensors | Cameras only | 13 cameras, 4 lidars, plus radar |
| Controls | No steering wheel or pedals | Robotaxi built on a Zeekr vehicle |
| Public rides began | 3 September 2026, Austin | Ojai rides from May 2026 |
| Scale | 45 Cybercabs registered in Texas in early September 2026 | About 500,000 paid rides a week (spring 2026) |
| Running cost (Ark’s estimate) | About 50% lower per mile | Baseline |
The scale gap is huge. InsideEVs reported that Waymo doubled its weekly ridership in under a year, and the company aims for one million paid rides a week by the end of 2026. Tesla is still counting its Cybercabs in the dozens or hundreds.
What could still go wrong?
Regulation is the big one. In early September 2026, just as public rides started, the National Highway Traffic Safety Administration opened an audit into whether Tesla correctly self-certified the Cybercab under federal safety standards. Those standards still assume a car has a steering wheel and pedals.
As Not a Tesla App reported, Tesla skipped the federal exemption route that other makers of cars without manual controls have used. It self-certified up to 1,000 Cybercabs instead. NHTSA sent 21 questions, and Tesla asked for time until 20 November. The agency gave it until 30 October 2026.
My view: the cost argument is plausible, but it’s an argument about 2030. Waymo is winning in 2026, and Tesla has a habit of missing autonomy deadlines. Rivals aren’t standing still either, as Rivian’s plans show in Rivian R4 Teased: Why Scaringe Says Autonomy Will Reshape Cars.
What happens next
- 30 October 2026: Tesla’s answers to NHTSA’s 21 questions are due. A tough response from the agency could slow the Cybercab rollout.
- Expansion: Dallas has been named as the next Cybercab market after Austin. Watch whether that launch happens without safety staff.
- Waymo’s target: If Waymo reaches one million weekly rides by the end of 2026, the scale gap gets wider.
- Federal rules: NHTSA’s proposal to drop manual-control requirements for driverless cars has not been finalised. If it is, Tesla’s path gets much easier.
FAQ
Who is Tasha Keeney?
She is director of investment analysis at Ark Invest and leads the firm’s research on autonomous technology and robotics.
Is the Tesla Cybercab carrying paying passengers?
Yes. Tesla began commercial Cybercab rides in parts of Austin, Texas, on 3 September 2026 through its Robotaxi app.
Why is NHTSA investigating the Cybercab?
The agency wants to know whether Tesla wrongly treated some federal safety standards as not applying, because the car has no steering wheel, pedals or mirrors.
Next step: watch for Tesla’s NHTSA filing after 30 October 2026. It will tell you more about the Cybercab’s near-term future than any price target.
This article is for information only and is not investment advice.
2 comments