A Tesla Model Y owner in Colorado says the vehicle’s Full Self-Driving (FSD) software drove his car directly into a lake, adding a dramatic and visually striking incident to the growing catalog of concerns surrounding Tesla’s advanced driver-assistance technology. The episode, which was captured in part by the vehicle’s own onboard cameras, has reignited a fierce debate about the safety, reliability, and regulatory oversight of autonomous driving systems — and whether Tesla’s aggressive rollout of FSD is outpacing the technology’s actual capabilities.
The incident, first reported by Futurism, involved a Tesla Model Y that was operating in Full Self-Driving (Supervised) mode when it allegedly veered off the road and plunged into a body of water in Colorado. The driver, who managed to escape the sinking vehicle, reported that FSD was actively engaged at the time of the incident. Images circulating on social media showed the white Model Y partially submerged, its front end beneath the waterline, offering a stark visual that quickly went viral across platforms including X (formerly Twitter).
What Happened at the Lake: A Closer Look at the Incident
According to reports, the Tesla owner engaged FSD while driving on a road near the lake. The system, which Tesla markets as capable of navigating complex driving scenarios including city streets, intersections, and highway merges, apparently failed to recognize the road’s trajectory or the presence of the body of water ahead. Rather than correcting course or alerting the driver to take over, the vehicle continued forward and entered the lake.
The driver has stated publicly that he was monitoring the system as Tesla instructs all FSD users to do, but that the car’s actions were sudden enough that intervention was not possible in time. Tesla’s FSD system requires drivers to remain attentive and keep their hands on the steering wheel, a stipulation the company uses to emphasize that FSD is not a fully autonomous system despite its name. Critics have long argued that the branding itself — “Full Self-Driving” — creates a dangerous expectation gap between what the technology promises and what it actually delivers.
Tesla’s FSD: A Technology Under Intensifying Scrutiny
Tesla’s Full Self-Driving software has been the subject of multiple federal investigations by the National Highway Traffic Safety Administration (NHTSA). The agency has previously opened probes into Tesla’s Autopilot and FSD systems following a series of crashes, some of them fatal. In 2024, NHTSA expanded its scrutiny after receiving hundreds of complaints about unexpected braking, failure to stop at intersections, and other erratic behaviors exhibited by vehicles running FSD.
Tesla CEO Elon Musk has repeatedly touted FSD as the company’s most transformative product, projecting that it would eventually enable a fleet of autonomous robotaxis that could generate enormous revenue. Musk has set — and missed — numerous deadlines for achieving full autonomy, dating back to 2016 when he first promised that all Tesla vehicles being produced had the hardware necessary for full self-driving capability. Despite these delays, Tesla has continued to expand the FSD beta program to hundreds of thousands of vehicles across North America, charging customers up to $12,000 for the software package or offering it as a $199-per-month subscription.
The Branding Problem: ‘Full Self-Driving’ That Isn’t Fully Self-Driving
The Colorado lake incident has once again thrown a spotlight on what consumer advocates and safety regulators have called Tesla’s misleading branding. The system is classified by the Society of Automotive Engineers as a Level 2 driver-assistance system, meaning the human driver is expected to remain fully engaged and ready to take control at any moment. By contrast, truly autonomous vehicles — such as those operated by Waymo, Alphabet’s self-driving subsidiary — operate at Level 4, meaning they can handle driving tasks without human intervention in defined areas.
“The name ‘Full Self-Driving’ is inherently deceptive,” said Philip Koopman, a Carnegie Mellon University professor who specializes in autonomous vehicle safety, in previous public statements on the matter. Consumer Reports, the Insurance Institute for Highway Safety, and multiple state attorneys general have raised similar objections. California’s Department of Motor Vehicles filed a complaint against Tesla in 2022, accusing the company of making “untrue or misleading” claims about FSD and Autopilot capabilities. Despite these challenges, Tesla has not changed the product’s name.
A Pattern of High-Profile FSD Failures
The lake plunge is far from an isolated event. Over the past several years, Tesla vehicles operating on Autopilot or FSD have been involved in a disturbing number of incidents. NHTSA has investigated dozens of crashes involving Tesla’s driver-assistance systems, including collisions with emergency vehicles parked on the side of the road, a recurring issue that led to a formal recall of more than two million vehicles in December 2023. That recall, one of the largest in Tesla’s history, was addressed through an over-the-air software update rather than a physical repair — a practice that itself has drawn regulatory questions about whether software patches are sufficient remedies for safety-critical defects.
In addition to the emergency vehicle collisions, there have been incidents involving Tesla vehicles running red lights, failing to detect pedestrians, and making erratic lane changes. A 2024 investigation by The Washington Post documented numerous cases in which Tesla drivers reported that FSD or Autopilot behaved unpredictably, sometimes in ways that could have been catastrophic. The pattern has led some safety experts to question whether Tesla’s approach to developing autonomous driving technology — which relies heavily on cameras and artificial intelligence rather than the lidar sensors used by most competitors — is fundamentally limited.
Tesla’s Vision-Only Approach: Innovation or Recklessness?
Tesla’s decision to rely exclusively on a camera-based vision system, which it calls Tesla Vision, has been one of the most debated strategic choices in the automotive industry. Musk has argued that cameras, combined with powerful neural networks, can replicate and eventually surpass human perception. He has dismissed lidar — the laser-based sensing technology used by Waymo, Cruise, and most other autonomous vehicle developers — as an expensive and unnecessary “crutch.”
However, incidents like the Colorado lake plunge raise questions about the limitations of a vision-only system. Cameras can be affected by glare, low light, adverse weather conditions, and unusual environmental configurations — such as a road that terminates at a body of water. Lidar systems, by contrast, create three-dimensional maps of the environment that are less susceptible to such visual ambiguities. While Tesla’s neural networks have improved dramatically with each software update, the company’s approach means that edge cases — rare but potentially dangerous scenarios that the AI has not been adequately trained on — remain a persistent vulnerability.
Regulatory and Legal Implications Mount
The incident is likely to add momentum to ongoing regulatory efforts aimed at imposing stricter oversight on advanced driver-assistance systems. In Congress, bipartisan legislation has been introduced that would require NHTSA to establish performance standards for automated driving systems and mandate more transparent reporting of crashes involving such technology. The European Union has already moved ahead with its own regulatory framework, requiring that Level 2 systems include more robust driver-monitoring technology and clearer labeling to prevent consumer confusion.
On the legal front, Tesla faces a growing number of lawsuits from crash victims and their families who allege that FSD and Autopilot are defective products. In several cases, plaintiffs have argued that Tesla’s marketing materials and Musk’s public statements created an unreasonable expectation that the vehicles could drive themselves safely. Tesla has generally defended itself by pointing to the terms of service that FSD users must agree to, which explicitly state that the driver must remain attentive at all times. But legal experts say that defense may become harder to sustain as evidence mounts that the system’s failures are not merely the result of driver inattention.
What Comes Next for Tesla and the Future of Autonomous Driving
For Tesla, the stakes could not be higher. The company’s valuation is heavily predicated on the promise of autonomous driving. Wall Street analysts have estimated that FSD and the eventual robotaxi business could account for a significant portion of Tesla’s future revenue — some bullish projections value the autonomous driving opportunity alone at hundreds of billions of dollars. Any regulatory action that limits FSD’s deployment, or a major legal judgment that holds Tesla liable for a crash caused by the software, could have profound financial consequences.
Meanwhile, competitors are making steady progress. Waymo has expanded its fully autonomous ride-hailing service to multiple cities, including San Francisco, Phoenix, and Los Angeles, accumulating millions of miles of driverless operation with a safety record that, while not perfect, has generally been well-received by regulators. The contrast between Waymo’s cautious, geographically limited rollout and Tesla’s mass-market deployment of a system that still requires human supervision is becoming increasingly stark.
The image of a Tesla Model Y sinking into a Colorado lake is, in many ways, a metaphor for the broader tensions surrounding autonomous driving technology: the collision between Silicon Valley’s move-fast ethos and the physical, life-and-death realities of operating multi-ton vehicles on public roads. As Tesla pushes forward with its vision of a self-driving future, incidents like this one serve as a sobering reminder that the gap between aspiration and execution remains dangerously wide.