For years, Tesla CEO Elon Musk has promised that fully autonomous vehicles were just around the corner. Meanwhile, Alphabet’s Waymo has been quietly racking up millions of paid robotaxi rides across multiple American cities — without a human behind the wheel. The divergence between the two companies’ approaches to autonomous driving has never been starker, and the data increasingly suggests that Waymo’s methodical, sensor-heavy strategy is pulling decisively ahead of Tesla’s camera-only, software-dependent bet.
As reported by Neural Foundry, the gap between the two companies isn’t merely philosophical — it’s operational. Waymo is already running a commercial robotaxi service, Waymo One, that completes over 150,000 paid trips per week across San Francisco, Phoenix, Los Angeles, and Austin. Tesla, by contrast, has yet to launch a single commercial autonomous ride without a human safety driver. The contrast raises uncomfortable questions for Tesla investors who have long priced in the promise of full self-driving as a major component of the company’s future valuation.
The Sensor War: Lidar and Cameras vs. Cameras Alone
At the heart of the technical divide is a fundamental disagreement about what machines need to see in order to drive safely. Waymo employs a multi-layered sensor suite that includes lidar (light detection and ranging), radar, and cameras working in concert. Lidar provides precise three-dimensional mapping of the vehicle’s surroundings, measuring distances with centimeter-level accuracy. Radar detects objects and their velocities even in poor weather conditions. Cameras supply visual context — reading traffic signs, identifying lane markings, and recognizing pedestrians.
Tesla, under Musk’s direction, famously abandoned radar sensors in 2021 and has committed entirely to a vision-only approach it calls “Tesla Vision.” The argument, as Musk has articulated repeatedly, is that humans drive with only two eyes and a brain, so a sufficiently advanced neural network processing camera feeds should be able to do the same. The problem, as Neural Foundry points out, is that this analogy breaks down in practice. Human vision is backed by decades of embodied experience, intuitive physics understanding, and a biological neural architecture that remains far more sophisticated than any artificial system currently deployed. Camera-only systems struggle with depth perception, glare, rain, fog, and the kind of edge cases — a white truck against a bright sky, for instance — that have been implicated in fatal Tesla Autopilot crashes.
Safety Records Tell a Diverging Story
The safety data compounds the technical argument. Waymo published a peer-reviewed safety study in late 2023 and early 2024 showing that its autonomous vehicles had an 85% lower rate of injury-causing crashes compared to human drivers over more than 22 million miles of autonomous driving. The company has been transparent about its disengagement reports filed with the California DMV, which show a steadily declining rate of situations where a human operator needs to take over.
Tesla’s Full Self-Driving (FSD) Beta, now rebranded as FSD (Supervised), still requires a human driver to remain attentive and ready to intervene at all times. The National Highway Traffic Safety Administration (NHTSA) has opened multiple investigations into Tesla’s Autopilot and FSD systems following fatal and serious crashes. In 2023, Tesla recalled over two million vehicles to update Autopilot software after NHTSA found the system’s driver monitoring was insufficient. The distinction is significant: Waymo’s cars operate with no human fallback in their operational domains, while Tesla’s system is explicitly classified as a Level 2 driver-assistance feature, meaning the human is always supposed to be in control.
The Business Model Divergence
Beyond technology, the two companies are pursuing fundamentally different business models. Waymo operates a ride-hailing service — essentially an autonomous Uber — where the company owns and operates the vehicles. Revenue comes from fares. The model requires enormous capital expenditure on vehicles, sensors, maintenance, and mapping, but it also means Waymo captures the full economic value of each ride. Alphabet has invested billions into Waymo, and the unit does not yet turn a profit, but the path to monetization is clear and already generating real revenue.
Tesla’s vision, as Musk has described it, is to turn every Tesla vehicle into a potential robotaxi through over-the-air software updates. Owners would be able to send their cars out to earn money autonomously when not in use, with Tesla taking a cut of the fare. It’s an extraordinarily ambitious concept — one that would, if realized, create a massive asset-light network overnight. But it depends entirely on achieving full autonomy through cameras and software alone, a milestone that has been perpetually “next year” since Musk first promised it in 2016. At Tesla’s Robotaxi unveiling event in October 2024, the company showed off a prototype vehicle called the Cybercab, but provided few concrete details on regulatory approval timelines or commercial launch dates.
Mapping, Geofencing, and the Scalability Question
One argument frequently made in Tesla’s favor is scalability. Waymo’s approach requires detailed pre-mapping of every area where its cars will operate, and the company geofences its vehicles to those mapped zones. Expanding to a new city takes months of preparation. Tesla’s system, in theory, should work anywhere because it relies on real-time visual processing rather than pre-built maps. This is a legitimate advantage — if the technology actually works at a safety level sufficient for unsupervised operation.
But Waymo’s expansion pace has been accelerating. The company announced plans to expand to Miami and additional cities through 2025, and its partnership with Uber in select markets has broadened its reach further. As Neural Foundry observes, Waymo’s geofenced approach may be slower to scale geographically, but it produces a service that actually works safely and commercially today. Tesla’s approach promises universal coverage but has not yet delivered a single market where unsupervised autonomous driving is available to customers.
The Valuation Disconnect
Perhaps the most striking aspect of the Waymo-Tesla comparison is the financial one. Tesla’s market capitalization — hovering near or above $800 billion through much of 2024 and into 2025 — bakes in enormous expectations for autonomous driving revenue. Analysts at Morgan Stanley and others have attributed hundreds of billions of dollars in Tesla’s valuation to the anticipated robotaxi network. Yet the company with the most advanced, commercially deployed autonomous driving technology in the United States is not Tesla but Waymo, which as a subsidiary of Alphabet doesn’t carry a separate public market valuation but has been estimated by analysts to be worth between $30 billion and $100 billion.
This creates what some analysts view as a significant mispricing. Tesla is valued as though it has already solved autonomy, while the company that has actually solved it — at least within defined operational domains — is valued at a fraction of that implied price. The disconnect is not lost on institutional investors, though Tesla’s stock price has historically been driven as much by narrative momentum and retail investor enthusiasm as by traditional valuation metrics.
Regulatory Headwinds and Tailwinds
Regulation adds another layer of complexity. Waymo has secured permits to operate commercially without safety drivers in multiple jurisdictions, a process that required extensive engagement with state and local regulators, public hearings, and the submission of detailed safety data. Tesla has not applied for — or received — any comparable permits for unsupervised autonomous vehicle operation.
The regulatory environment is evolving rapidly. California, Arizona, and Texas have established frameworks for autonomous vehicle deployment, and federal legislation has been discussed but not yet passed. Waymo’s years of regulatory engagement give it a significant head start. Tesla, which would need to demonstrate that its camera-only system meets safety thresholds for driverless operation, faces a longer road to regulatory approval — assuming the technology reaches that level of performance.
What the Next Two Years Will Reveal
The period from 2025 through 2027 will likely be decisive. Tesla has indicated it plans to launch a supervised robotaxi service in Austin, Texas in June 2025, though details remain sparse and the “supervised” qualifier is notable — it suggests a human will still be present. Musk has said unsupervised FSD could follow, but timelines from Tesla on this subject have historically proven unreliable.
Waymo, meanwhile, continues to expand its operational footprint, increase its weekly ride count, and accumulate the kind of real-world safety data that regulators and the public demand. The company recently began testing in new weather conditions and more complex urban environments, pushing the boundaries of where its vehicles can operate.
For industry observers, the lesson is becoming increasingly clear: the company that chose the harder, more expensive, more methodical path to autonomy is the one that got there first. Whether Tesla can close the gap — or whether its camera-only approach has a fundamental ceiling — remains the most consequential open question in the autonomous vehicle industry. The market, for now, is betting on Musk. The roads, however, are telling a different story.