We take a look at the self-driving wars. Breaking down the current state of the market, the technology each player is betting on, and how we think the industry will end up.

Waymo and Tesla are battling for supremacy in the emerging self-driving car market.

Waymo, owned by Google parent Alphabet, is leveraging a sensor-heavy approach to self-driving; retrofitting vehicles with a barrage of large and obtrusive sensors: 13+ cameras, 4 LiDAR, and 6 radar, to be exact.

Tesla, by contrast, is doubling down on a vision-only, software-first strategy, with 8 external cameras, having phased out the use of radar and ultrasonic sensors since 2023.

Which approach will yield the best results? Which will be the first company to build a sustainable, large-scale autonomous ride-hailing service?

Both companies face complex regulatory hurdles, safety and liability questions, and immense operational costs as they race to the future of autonomous transport.

Levels of Autonomy

The Society of Automotive Engineers (SAE) International has established a widely recognized classification system that defines size levels of driving automation for on-road vehicles:

0. No Automation, but can include lane departure warnings and emergency braking alerts

1. Driver Assistance: can assist with either steering or speed controls, but not both simultaneously

2. Partial Driving Automation: can assist with both steering and speed controls, under certain conditions

3. Conditional Driving Automation: the system manages all aspects of driving under specific conditions

4. High Driving Automation: the system performs all driving tasks within defined geographic areas

5. Full Driving Automation: the system can perform all driving tasks under all conditions

Waymo

Waymo launched in 2009 as Google X’s self-driving project.

(Waymo's roots actually trace back to the Stanford Racing Team, which won the 2005 DARPA Grand Challenge with its autonomous vehicle, Stanley.)

In 2015, Waymo achieved a significant milestone by completing the world's first fully driverless ride on public roads in Austin, Texas, with a legally blind passenger, Steve Mahan. Waymo then became an independent subsidiary under Alphabet Inc. in December 2016.

Waymo adds pace

As of writing, Waymo operates over 250,000 paid trips per week across San Francisco, Phoenix, Los Angeles, and Austin, totaling more than 4 million trips in 2024 (Reuters).

It reports 56.7 million autonomous, driverless miles logged on public roads, demonstrating Level 4 autonomy within geofenced zones.

In early May 2025, Waymo announced a partnership with Magna International to build a new factory in Mesa, Arizona; the mission is to produce over 2,000 Waymo-equipped Jaguar I-PACE vehicles next year.

Tesla

Tesla introduced its Autopilot system in 2014, initially incorporating a combination of cameras, radar, and ultrasonic sensors.

In 2016, the company began equipping all new vehicles with hardware it claimed was capable of full autonomy, marking the start of its Full Self-Driving (FSD) journey.

Despite CEO Elon Musk's notoriously bold predictions of achieving full autonomy by 2018, the rollout has been gradual and iterative.

A significant milestone occurred in October 2020 with the release of the FSD Beta, allowing select users, filtered based on their driving 'safety score', determined by Tesla, to test advanced features like city street navigation. By November 2022, the beta program expanded to all North American owners who had purchased the FSD package, regardless of their safety scores.

It is important to note that all Tesla FSD currently operates at SAE Level 2, meaning it requires active supervision by the driver.

Despite safety controversies and limited transparency around intervention rates, Tesla continues to roll out incremental software updates—most recently the Spring Update to FSD early access testers—with the promise of a fully unsupervised FSD v14 on the horizon.

Autonomous Austin dreams

Tesla plans to launch its first robotaxi service in Austin, Texas, in summer 2025, supported by roughly 300 test operators and an early-access app for Autopilot engineers.

Side note: Tesla's 'Robotaxi' trademark request was just rejected by the US Patent Office at the time of writing

Different technology approaches

Waymo

Waymo equips OEM-manufactured vehicles with a system of high-resolution lidar, radar, and multiple cameras to perceive the environment with centimeter-level accuracy.

This multilayered sensor array allows Waymo to detect objects in all lighting and weather conditions, enabling robust Level 4 operations within specifically pre-mapped zones.

The data from the sensor suite feeds into Waymo's AI-driven perception system, which identifies and tracks objects such as pedestrians, cyclists, vehicles, and road signs. Waymo employs advanced machine learning models, including its proprietary VectorNet, a graph neural network that predicts the trajectories of surrounding objects, enabling the vehicle to make informed driving decisions.

Waymo utilizes a simulation platform known as Carcraft, where virtual vehicles navigate through detailed models of real-world environments. This allows Waymo to test and refine its driving algorithms across countless scenarios, including rare and hazardous situations.

Style and speed problem

Waymo's current-generation sensors are not just visually obtrusive, but also significantly reduce the aerodynamics of the vehicle they are attached to. Waymo is working hard to reduce the size of its sensor systems and to design kits that seamlessly blend into the existing body of the vehicle.

Tesla

Tesla’s approach skips lidar entirely, relying on eight surround-view cameras. Tesla’s custom on-board FSD computer then interprets raw image data. While this setup yields a simpler hardware stack, it can struggle in low-visibility scenarios and depends heavily on neural network generalization.

Musk has been adamant for some time that if a human can drive with just visual input, so too should cars.

Orders of magnitude more training data

Tesla has racked up over 3.6 billion miles driven on its Full Self-Driving beta, though only at Level 2 autonomy.

If the mission is general (Level 5) autonomy, can Tesla's broader approach to self-assisted driving help it get there quicker than Waymo?

It can be analogized that Tesla is going for a breadth-first search solution, collecting as much driver-assisted real-world data as possible, whilst Waymo is attempting a depth-first search approach by perfecting very specific geographic areas at once.

Different business models

Waymo

Waymo’s core offering, Waymo One, is a ride-hailing service charging fares comparable to Uber and Lyft.

Financial analysts at Morgan Stanley project Waymo could generate over $2.5 billion in annual revenue by 2030, driven by fleet expansion and licensing deals (Investor's Business Daily).

Beyond direct rides, Waymo licenses its Driver software to automakers like Jaguar, and has announced exploratory autonomous vehicle projects with Toyota (The Verge).

The incremental cost of lidar has fallen, but it remains a significant component of unit economics. High volume production is essential for Waymo to come down the cost curve to help build a sustainable business. Although with the purse strings of Alphabet at its disposal, it is hardly a startup desperate to prove profitability.

Tesla

Tesla’s FSD package retails for up to $15,000 per vehicle, with customers paying upfront or via subscription, providing a high-margin software revenue stream that has helped cushion margins amid slowing car sales across the US and Europe.

Looking ahead, Tesla intends to pool its fleet into a robotaxi network, with Elon Musk envisioning a shared-economy model where owners can “rent out” their vehicles when idle.

Above: concept images of the proposed Tesla robotaxi

The Austin Showdown

Austin is shaping up as the first meaningful battleground: Waymo has been operating since March, while Tesla readies its robotaxi fleet for a June debut.

Waymo’s cautious, sensor-rich model contrasts with Tesla’s software-only rollout, setting up a real-world test of Level 4 versus Level 2+ offerings.

Consumer adoption and regulatory feedback here will likely influence each company’s next expansion choices.

Challenges Ahead

Regulatory and Safety Hurdles

Only a handful of jurisdictions permit true Level 4 services, and both companies must navigate patchwork regulations.

Waymo’s geo-fenced operations simplify approvals, but expansion into denser urban areas will trigger fresh reviews.

Tesla’s planned unsupervised FSD will face intense scrutiny given its history of requiring driver intervention.

Liability and Public Trust

High-profile crashes involving Tesla’s Autopilot have fueled skepticism, while Waymo’s spotless safety record is built on conservative operational limits.

Yet the public remains wary of handing over control, particularly in shared-ride scenarios.

Clear liability frameworks and transparent safety data (intervention rates, disengagement statistics, etc.) will be critical to building trust.

Capital Intensity and Unit Economics

Building a domestic factory in Arizona is a multihundred-million-dollar commitment for Waymo, and margins are razor-thin at low volume.

Tesla’s hardware-light approach saves capital upfront but may incur hidden costs if vision-only fails in edge cases, leading to expensive retrofits.

Both must still achieve significant scale to cover R&D and regulatory compliance investments.

Looking Ahead

Over the next 12 months, the industry will watch four key milestones:

1. Waymo’s factory ramp-up in Mesa and its ability to deliver 2,000 new autonomous Jaguars by year-end.

2. Tesla’s June robotaxi launch in Austin—will it operate reliably without safety driver oversight?.

3. Geographic expansion beyond initial cities, testing each company’s ability to adapt to new traffic patterns and regulations.

4. Financial inflection points, as both firms seek to demonstrate a clear path to profitability through rides and software subscriptions.

The self-driving car wars between Waymo and Tesla represent more than a technical arms race; they embody contrasting visions of autonomy’s future.

Will high-fidelity sensor kits and controlled rollouts win the day, or will mass-market vision AI and software agility prevail?

The car market is enormous - in 2024, the global vehicle market value was $2.84 trillion (The Business Research), and the ride-hailing market reached $173.6 billion (P&S Intelligence).

There is a tendency to believe there will be space for both Tesla and Waymo as the market evolves. However, autonomous driving could be different. There may be natural monopoly benefits of having a wider distribution of your autonomous fleet. For example, imagine if every vehicle on the road were part of the same fleet; they could all coordinate as an effective hive mind, helping to seamlessly operate traffic flows and navigate around each other safely.

With such a valuable prize at stake, it will be interesting to see how this battles develops as it intensifies.

— Alex BB