The past several months of reporting on Alphabet's autonomous driving subsidiary Waymo reveal a company at a familiar inflection point — one I recognize from the early days of the horseless carriage. The fundamental engineering challenge of making a vehicle drive itself is, in many respects, solved. What remains is the harder, more painstaking work: making that capability reliable across the infinite edge cases of real-world operation, navigating a fragmented regulatory landscape, and managing public perception in an environment where one failure can undo years of patient progress.
The claims coalesce around several reinforcing narratives: a distinctive crash profile dominated by stationary-vehicle incidents, recurring operational failures in non-standard scenarios, uneven political reception across U.S. cities, and a competitive landscape where rivals operate under meaningfully different constraints and risk profiles. For Alphabet investors, these patterns collectively suggest that Waymo's primary challenges are shifting from core perception and decision-making — the hard problems of the last decade — to operational resilience, regulatory navigation, and risk management. Each carries material implications for deployment timelines, capital requirements, and the ultimate shape of the autonomous vehicle market.
The Stationary-Vehicle Safety Paradox
The single most corroborated finding across this dataset is that Waymo vehicles are disproportionately involved in crashes while stationary — a pattern that demands careful parsing, for it reveals as much about Waymo's engineering philosophy as it does about its vulnerabilities.
Multiple independent analyses of NHTSA Standing General Order data converge on this finding. Of Waymo's 693 reported crashes in the SGO dataset, 59% (411 incidents) occurred at zero pre-crash speed — the Waymo vehicle was stationary when impacted 2. Among pick-up/drop-off crashes specifically, 64 of 76 (84%) happened at 0 mph 2, representing 9% of total reported crashes 2. NHTSA reports show the proportion of Waymo-involved crashes where the vehicle was parked or stopped prior to impact rose from 46% in 2022–2023 to 65% in 2024, and further to 70% in 2025 37 — though the year-over-year increase slowed after 2024 37.
The qualitative descriptions of these incidents are instructive. Waymo vehicles are frequently rear-ended at intersections, struck by vehicles reversing or rolling back into them, sideswiped, or hit by vehicles backing out of driveways 31. In one particularly severe case — the only collision in the February-March 2026 NHTSA reports requiring hospitalization — a Waymo vehicle parked at a curb was hit by a vehicle that ran a stop sign, striking both a passenger inside and a passenger standing outside with the door open 31. Another incident involved a motorcyclist who rear-ended a Waymo, was thrown into the next lane, and fatally struck by a human-driven vehicle 31. In Austin, a human driver rear-ended a Waymo, drove around it, and then reversed into its front 2.
From an engineering perspective, this pattern cuts both ways. It suggests that Waymo's defensive driving logic — its systematic tendency to stop and wait rather than assert into traffic — keeps it from being the proximate cause of most collisions. Waymo itself studied all human-driver fatalities in its Phoenix service area and concluded its vehicles could have prevented most of those events had they been the not-at-fault car, and all of them had they been the at-fault car 31. This is the behavior one would expect from a system designed from first principles to prioritize safety over traffic flow.
However, the stationary-vehicle profile also exposes a structural vulnerability. Waymo's operating model creates situations where human drivers — unaccustomed to vehicle behavior that prioritizes caution over assertiveness — collide with predictable regularity. This is less a technology failure than a fleet-operations and risk-management challenge, but it carries real safety consequences nonetheless. The insurance and liability implications for Alphabet are material: a growing volume of third-party collision costs that software improvements alone cannot eliminate. Much as early automobile owners bore the cost of accidents caused by horses spooking at the unfamiliar sight of a motorcar, Waymo is bearing costs generated not by its own engineering failures, but by the gap between its operating logic and human expectations.
Operational Edge Cases: The Problem of Non-Standard Scenarios
A second major theme concerns Waymo's handling of scenarios that fall outside its operational design domain — the edge cases that, by definition, are rare but collectively define whether a system is truly production-ready.
The San Francisco Fire Department Chief, Patrick Rabbitt, publicly stated that Waymo vehicles' default behavior is to "freeze" when encountering obstacles, causing them to stop near fire stations 8 — a problem he described as occurring "frequently" 8 and one that prevents fire trucks from responding to emergencies in a "timely and appropriate" manner 8. Reports indicate Waymo vehicles have blocked fire station access in San Francisco, with accounts suggesting the problem is worsening over time 8. One analysis characterizes this behavior as suggesting inadequate handling of non-standard situations 8.
A power outage in San Francisco caused Waymo vehicles to stop without warning — a realized system failure scenario that should give any engineer pause 34. Waymo subsequently updated its response systems 34, but the incident underscores the vulnerability of autonomous fleets to infrastructure disruptions, a class of risk that lies entirely outside the vehicle's perception and decision-making systems.
The school-zone incidents are particularly concerning from a safety engineering standpoint. A Waymo vehicle was recorded driving the wrong way in an Alamo Heights Independent School District school zone in San Antonio, Texas 19. Waymo acknowledged the incident indicated the need for technical fixes to its mapping, navigation, or decision-making algorithms 19. NHTSA has conducted investigations of Waymo vehicle behavior near school zones 34, and incidents in which pedestrians were struck by Waymo robotaxis occurred in or near school zones or near school buses — locations with vulnerable populations such as children 11. One analysis suggests that if Waymo's systems fail to detect and avoid pedestrians in school zones, it indicates a technology or perception system deficiency 11. This is precisely the sort of finding that, in the early automobile era, would have prompted a fundamental reexamination of braking systems and pedestrian detection — and it deserves equivalent engineering seriousness today.
Other edge cases accumulate: a Waymo vehicle navigating incorrectly into a drive-thru lane in San Antonio 12; a vehicle failing to navigate a routine urban scenario involving trucks parked in a loading zone without external intervention 18; a vehicle becoming stranded on a narrow downtown street 18; an empty Waymo vehicle being swept into a flooded roadway in San Antonio, leading to a temporary service pause 10. In Nashville, multiple incidents involving Waymo vehicles stopping in intersections were reported across different locations 15. A San Antonio vehicle did not respond appropriately to a police officer, who was described as "helplessly" waving his arms at the vehicle 12; two police officers responded to the scene of a stranded Waymo vehicle, taking over 10 minutes to reach someone capable of moving it 18. The stranded vehicle had the potential to block emergency vehicles such as fire trucks and ambulances 18.
A notable operational quirk: multiple Waymo vehicles were observed honking at each other repeatedly throughout the night in San Francisco while attempting automated parking maneuvers 20, an edge case that illustrates how low-probability scenarios can cascade into broader coordination failures 20. Municipal regulations for autonomous vehicle testing may be affected by quality-of-life complaints such as these 20 — a reminder that public tolerance for autonomous systems is shaped not only by safety, but by the more mundane irritations of coexistence.
The Regulatory Patchwork: A Structural Headwind
From my perspective as someone who watched the early automobile industry negotiate its place in a world built for horses, the most critical structural insight from this dataset is clear: there is no federal preemption for autonomous vehicle standards, meaning localities retain significant power to set or block autonomous vehicle operations 35. This creates a fragmented regulatory environment where Waymo must navigate individual city and state requirements — a pattern that historically has slowed adoption and raised costs for every transportation innovation.
New York City presents a clear example. Mayor Zohran Mamdani declined to answer whether his administration would extend Waymo's testing license 25, and Waymo faced political opposition from his administration 25. The city can prohibit specific autonomous vehicle operators from operating within its jurisdiction 35, and one account suggests New York City could be blocked for autonomous vehicle operations during the current mayor's term 28. At the state level, Governor Kathy Hochul paused legislation that would have allowed driverless vehicle services to operate across the rest of New York State 29.
In Portland, Oregon, reactions to Waymo's entry are mixed: some residents express excitement while others express concern about safety and impacts on rideshare drivers 36. Residents and advocacy groups have raised safety concerns 36, and Tom Stenson of Disability Rights Oregon spoke publicly about the safety and accessibility implications of autonomous vehicles for people with disabilities 36. Some Portland city leaders raised concerns about potential safety risks 9.
At the federal level, Senator Ed Markey (D-MA) released a report revealing that seven leading autonomous vehicle companies declined to disclose remote assistance usage data 7 and urged NHTSA to investigate this refusal 7. Separately, NHTSA initiated a probe into Tesla's Actually Smart Summon feature in January 2025 following reports of minor collisions 37,38, and said it may reopen the investigation if warranted 38 — though the agency closed the probe on the basis that Tesla had already recalled the feature 38.
The absence of federal preemption means that Waymo's expansion trajectory will be lumpy, unpredictable, and expensive. Cities like Boston, Detroit, and Minneapolis face combined weather and infrastructure challenges that may keep them out of reach for years. This limits Waymo's near-term addressable market and raises the cost of market entry — a structural headwind that no amount of engineering excellence can eliminate.
Remote Assistance: The Undisclosed Variable
The role of remote human intervention in autonomous operations emerges as a significant but opaque factor — one that, from an engineering transparency standpoint, deserves far more scrutiny than it has received.
Senator Markey's report highlighted that seven AV companies refused to disclose remote assistance usage data 7. Waymo itself has a "substantial share" of overseas remote agents holding Philippine driver's licenses rather than U.S. licenses 26. Tesla acknowledged that remote staff can take full control of vehicles in certain operational scenarios 7, with direct input described as a "last resort" of limited duration used to move vehicles out of compromising positions without waiting for a field representative 26. Tesla stated that remote human driving interventions occur at speeds below 10 miles per hour 22,26.
In a notable Waymo incident in Austin, a vehicle drove past a school bus displaying an extended stop sign after receiving incorrect information from a remote assistant 26 — suggesting that remote assistance, intended as a safety backup, can itself introduce new failure modes. This is a fundamental systems engineering insight: every layer of redundancy introduces its own failure modes, and those failures must be validated with the same rigor applied to the primary system.
The governance and transparency implications are significant. If remote human intervention is more extensive than publicly acknowledged — and the refusal of seven companies to disclose usage data suggests it may be — then Waymo's "autonomous" operations are less automated than they appear. This matters for valuation, for regulatory classification, and for public trust. The early automobile industry learned this lesson the hard way: when the public discovers that a system is less capable than advertised, the resulting backlash can set progress back by years.
Competitive Landscape: Divergent Engineering Philosophies
The claims allow for useful comparisons across autonomous vehicle operators, revealing a landscape where different engineering philosophies and deployment strategies produce markedly different risk profiles.
Waymo reported 693 accidents in NHTSA SGO crash reports during the analyzed period 2, with an average police-reported crash rate of 1 crash every 500,000 miles driven 2. By comparison, Tesla reported 15 crashes across its autonomous vehicle operations 2, though Tesla's operational footprint and reporting standards differ significantly. Tesla's unsupervised operations are suspended during rain 31, and the last two periods had half as many active Tesla autonomous vehicles as in prior periods 31. Local ride spotters estimated 2-3 concurrent Tesla vehicles were operating in unsupervised mode in south Austin 31. A judge ruled that Tesla's use of "Autopilot" and "Full Self-Driving" in its marketing was misleading and violated state law 24 — a legal finding that echoes the historical pattern of overpromising undermining legitimate engineering progress.
Avride reported 36 crashes across its operations 2. Notably, Avride operates with safety operators who can disengage autonomous mode 31, and two of Avride's collisions occurred after autonomous vehicles stopped in intersections, safety operators disengaged autonomous mode, and then drove in reverse into vehicles behind them 31. In one Avride incident, the human safety operator behind the wheel did not intervene to prevent a collision with a duck — and the vehicle did not stop after the strike 5,6 — raising questions about the effectiveness of human supervision 6. Avride modified its testing area following the duck incident, excluding certain streets near the Mueller Lake neighborhood of Austin 6.
May Mobility reported 11 crashes and did not operate in Austin during the analysis period 2. May Mobility reported that its planner predicted a collision but the vehicle lacked honk and reverse capabilities to avoid it 2 — a limitation that, from a first-principles engineering perspective, represents a fundamental gap in actuation capability that should have been addressed before deployment.
Motional had a possible road-rage incident after one of its vehicles maneuvered around a stopped vehicle 31, and another Motional vehicle was struck by an object thrown by a pedestrian 31. Cruise, the General Motors subsidiary, experienced a major operational failure in October 2023 that led to suspension of its autonomous operations 21, a remote-operation failure in San Francisco in 2023 37, and reportedly failed in an attempt to develop custom silicon for its autonomous vehicles 32.
Baidu's Apollo Go robotaxis in China obstructed traffic and stranded passengers during a Wuhan incident 4, and Chinese authorities subsequently suspended autonomous vehicle licenses for Baidu 23. Baidu has been operating a fully driverless paid ride-hailing service in Wuhan 33.
A brief but telling signal: Sony terminated its Afeela electric vehicle collaboration with Honda, removing automotive manufacturing diversification from its operations 1. This indicates that even well-capitalized technology companies are reassessing their autonomous vehicle commitments — a sign that the market is developing more slowly than early projections anticipated.
Expansion Strategy: Where Geography Meets Economics
Waymo's geographic expansion appears deliberate and constrained — consistent with an engineering organization that understands the gap between technical feasibility and operational viability.
California expansion is described as a "gold rush opportunity" connecting cities between California locations 28. San Diego and Sacramento are considered near-certainties for expansion as they connect existing California operations 28, and three geographies — Texas, Florida, and California — are expected to be connected for autonomous vehicle operations 28. In San Diego, depot locations appeared to be finalized even during the testing phase 28.
However, Detroit, Denver, Minneapolis, Philadelphia, Pittsburgh, and St. Louis are cities affected by winter weather concerns and could be scratched from near-term expansion plans 28. Seattle is considered workable because cold weather is infrequent enough that operations can be suspended during winter weather events 28. Boston's driving conditions — narrow streets, aggressive drivers, snow and ice, and complex traffic patterns — present significant operational challenges 13, though a Waymo-branded vehicle was photographed there 13. The University at Buffalo Amherst campus serves as Waymo's base for winter weather testing 27.
The primary barrier to rural and suburban operations is cited as economics rather than technology 28, and Waymo has characterized coast-to-coast operation as technically feasible but not yet economically viable 28. This is a crucial distinction that I suspect is underappreciated by market observers: capital efficiency and per-mile economics — not engineering capability — are the binding constraints on Waymo's growth. The technology can navigate from New York to Los Angeles; the question is whether it can do so at a cost that generates a return on the capital invested.
Technical Capabilities and Data Assets
Despite the operational challenges documented above, several claims highlight Waymo's genuine technical strengths — capabilities that represent real competitive advantages and, in some cases, unexpected synergies within Alphabet.
Waymo vehicles transmit road defect data in real time to transportation authorities and drivers 16, and pothole detection data is transmitted to the U.S. Department of Transportation 16. Waymo provides its road-infrastructure pothole data to Waze, Alphabet's navigation platform, enabling drivers to avoid potholes and city planners to access aggregated data for road maintenance 14 — an example of intra-Alphabet data synergies that demonstrates how autonomous vehicle operations can generate value beyond the core transportation service.
Waymo reports nearly all of its SGO dataset accidents as occurring on "5th Generation ADS, Version 10" software 2, suggesting the dataset reflects a relatively homogeneous technical baseline — a favorable condition for systematic debugging and iterative improvement. Commenters noted that Waymo's internal operations consume significant Google Cloud Platform server capacity 3, representing both a cost center and an internal customer for Alphabet's cloud business.
Tesla claims a training data advantage for autonomous driving that is 35 times greater than that of Waymo 30, though this claim is self-reported and unverified. Tesla has not applied for a $3,000 California autonomous vehicle operating permit in over seven years, citing high regulatory hurdles 28.
Market Perception and User Behavior
Despite the operational challenges documented above, user behavior suggests strong demand — a finding that should give regulators and operators reason to proceed carefully, rather than dismiss public safety concerns.
In San Francisco, riders often choose Waymo even when Uber or a classic taxi would have been cheaper, citing a desire to avoid erratic human driving behavior or unwanted conversation 25. Some San Francisco users reported paying extra for Waymo rides despite higher prices because they perceived them as safer 30. Organic social media posts using a Waymo-branded hashtag have generated unpaid marketing exposure for Waymo 17.
However, one passenger reported that Waymo autonomous rides exhibit "phantom braking" that was "pretty startling" 30 — a reminder that the user experience, while apparently valued enough to command premium pricing, is not without friction.
Analysis and Significance
Collectively, these claims paint a picture of Waymo as a technology leader whose primary challenges are shifting from fundamental engineering to operational resilience, regulatory navigation, and public perception management. Several implications for Alphabet are worth examining from first principles.
First, the stationary-crash profile is a double-edged sword for liability and insurance. Waymo's defensive driving logic likely reduces its at-fault crash rate, which is favorable for liability exposure and regulatory comfort. However, the 65–70% stationary-crash share means that a large and growing portion of Waymo's incident burden is caused by human drivers hitting stationary Waymo vehicles. This creates a fleet operating cost — vehicle damage, downtime, towing, insurance claims — that is largely outside Waymo's control to eliminate through better software. The insurance and risk-transfer implications for Alphabet are material and, I suspect, underappreciated by the market.
Second, regulatory fragmentation is a structural headwind that will not resolve quickly. With no federal preemption in place, Waymo must negotiate city by city, state by state. New York's political opposition, Portland's mixed reception, and the patchwork of state-level permitting requirements create a lumpy, unpredictable expansion trajectory. Cities like Boston, Detroit, and Minneapolis may remain out of reach for years due to combined weather and infrastructure challenges. This limits Waymo's addressable market in the near to medium term and raises the cost of market entry.
Third, the remote assistance disclosures introduce a governance and transparency risk. Senator Markey's investigation, combined with the fact that a substantial share of Waymo's overseas remote agents hold Philippine rather than U.S. licenses 26, raises questions about quality control, training standards, and liability. If remote human intervention is more extensive than publicly acknowledged, Waymo's autonomous operations are less automated than they appear — a fact that matters for valuation, regulation, and public trust. Historical precedent suggests that when the public discovers such gaps, the resulting backlash can be severe.
Fourth, the competitive landscape remains bifurcated but favors Waymo in operational scale. Waymo leads in demonstrated miles and regulatory engagement. Tesla has a claimed data advantage but faces regulatory hurdles, a court ruling against its marketing claims, and a much smaller unsupervised operational footprint. Cruise is effectively sidelined. Baidu faces regulatory headwinds in China. Avride and May Mobility operate at much smaller scale. This suggests Waymo's competitive moat is real but not unassailable — operational scale begets data, which begets better technology, but regulatory and operational friction could slow the flywheel.
Fifth, Alphabet's role as both operator and cloud provider creates internal dynamics worth monitoring. Waymo consumes significant Google Cloud capacity 3, making it an important internal customer at a time when GCP is competing aggressively for external AI workloads. The data-sharing arrangement between Waymo and Waze 14 demonstrates internal synergy, but the net financial contribution of Waymo to Alphabet — factoring in GCP consumption, R&D spend, and liability reserves — remains opaque.
Key Takeaways
From the perspective of someone who has watched transportation revolutions unfold before, the lessons here are clear — and they are lessons about patience, transparency, and the gap between engineering capability and operational reality.
-
Waymo's crash profile is distinctive and structurally driven by human-driver behavior, not autonomous system failures. The fact that 65–70% of collisions involve a stationary Waymo vehicle suggests the technology is highly defensive, but also that Alphabet is bearing a growing volume of third-party collision costs that software improvements cannot eliminate. Investors should monitor insurance costs, vehicle downtime ratios, and fleet maintenance expenses as leading indicators of operating leverage.
-
Regulatory fragmentation is the primary near-term constraint on expansion, not technology. Without federal preemption, Waymo faces a city-by-city approval process where local political dynamics can delay or block deployment. The set of near-term viable expansion cities appears limited to Sun Belt and West Coast geographies, with winter-weather cities likely 3–5 years out.
-
Remote assistance transparency is an emerging governance risk. Senator Markey's investigation and the disclosure of overseas, non-U.S.-licensed remote operators raise questions about how much of Waymo's autonomous operations actually rely on human judgment from operators who may lack familiarity with U.S. driving conditions. This could become a focus for further regulatory scrutiny and could affect public trust.
-
Waymo's competitive position is strong but not without challenges. It leads in operational scale and demonstrated miles, but Tesla's claimed data advantage, the capital intensity of fleet expansion, and the uncertain timeline to positive unit economics mean that Waymo's path to profitability is longer and more capital-intensive than a pure technology-advantage narrative would suggest. The Sony-Honda collaboration termination 1 and Cruise's struggles 21,32 indicate that even well-resourced competitors are retrenching — a dynamic that is favorable for Waymo's market position but also signals that this market is developing more slowly than early projections anticipated.
The fundamental engineering challenge of autonomous driving — can a vehicle navigate from point A to point B without human intervention? — has been answered. The questions that remain are the ones that have always determined whether a technological innovation transforms transportation or remains a curiosity: Can it do so reliably across all conditions? Can it do so at a cost that generates economic return? And can it earn and maintain the trust of the public and the regulators who represent them? These are the questions that will determine Waymo's trajectory, and they are far from settled.
Sources
1. Japanese investments when EU bans US companies - fujitsu and others - 2026-04-11
2. Comparing pre-crash speeds between US ADS operators - 2026-04-24
3. Are hyperscalers turning into a winner take most market? Should I buy more $GOOGL or diversify? - 2026-04-29
4. China suspends new autonomous vehicle licenses following a traffic disruption involving Baidu Apollo... - 2026-04-29
5. Avride's self-driving car in Austin strikes and kills a beloved local duck, sparking community conce... - 2026-04-09
6. Self-driving car kills duck in Texas neighborhood, raising concerns about autonomous vehicles 🤖 IA:... - 2026-04-09
7. Senator Ed Markey Presses Autonomous Vehicle Companies Over Secrecy on Remote Assistance Practices ... - 2026-04-05
8. #Waymo is frequently now blocking our fire stations from access,” added Chief Patrick Rabbitt, the h... - 2026-04-30
9. Self-driving vehicle company #Waymo will launch fleet in #Portland. [Link] Self-driving vehicle com... - 2026-04-28
10. #Waymo #Self-DrivingCar #DroveItselfIntoFlood #SanAntonio #Texas San Antonio, Texas: news4sananton... - 2026-04-22
11. Girard Sharp is investigating potential claims for pedestrians struck by a Waymo robotaxi in a schoo... - 2026-04-21
12. Personally, I think the CEOs of these automated, driverless vehicles should be directly charged with... - 2026-04-16
13. [My robot overlords are finally arriving. #waymo #BostonMA Image: Parked white two door car, side ... - 2026-04-13
14. Waymo's robotaxi fleet will now use its sensors to detect and map potholes, sharing the data with Wa... - 2026-04-12
15. What could go wrong! #Waymo #Nashvegas #Fox17 [Link] Days after launch, Waymo vehicles block traffi... - 2026-04-12
16. Through the Waze for Cities program, Waymo vehicles will detect potholes and send ... - 2026-04-10
17. Six months ago, I said you’d have to be crazy to ride in a driverless car. Today, I took my first W... - 2026-04-05
18. Yesterday, an empty Waymo got stranded on a narrow street downtown, "confused" by trucks in a loadin... - 2026-04-04
19. It is only a matter of time before someone is killed by #Waymo. A #SanAntonio resident recorded a Wa... - 2026-04-03
20. This Eddy Burbank video features a clip from a story out of #SanFrancisco with #Waymo driverless car... - 2026-04-03
21. methodical vs. move-fast. one of these plays wins and it probably won't be the one everyone's bettin... - 2026-04-03
22. #Tesla Admits Its #Robotaxis Are Sometimes Driven by Remote Humans https://www.wired.com/story/tesl... - 2026-04-02
23. 2026-04-29 Briefing - alobbs.com - 2026-04-29
24. Most people still don’t want anything to do with robotaxis - 2026-04-15
25. Waymo to launch pilot program in London soon, full robotaxi service still this year - 2026-04-07
26. Robotaxi companies won’t say how often remote operators intervene - 2026-04-06
27. Waymo in NYC ? - 2026-04-18
28. Which cities are legally plausible next? - 2026-04-24
29. With Waymo Testing Halted, We Have A Rare Chance To Get Ahead of the ‘Driverless Revolution’ - 2026-04-08
30. Waymo starting to lose the self-driving cars race - 2026-04-24
31. NHTSA's April 2026 update of Autonomous Driving System incident reports - 2026-04-18
32. NVIDIA Doesn’t Matter (for Driving Automation) by Andrew Miller - 2026-05-01
33. Pony AI deploys driverless robotaxis in Dubai, plans commercial service launch in 2026 - 2026-04-20
34. Waymo Robotaxi Expansion: Autonomous Rides Launch in 4 New US Cities - 2026-04-30
35. The federal government should "preempt" cities and states to set standards for autonomous vehicles, ... - 2026-04-17
36. Waymo self-driving cars could come to Portland, Disability Rights Oregon talks safety - 2026-05-01
37. Recent developments of automated vehicles and local policy implications - npj Sustainable Mobility and Transport - 2026-04-27
38. NHTSA Ends Tesla ‘Smart Summon’ Probe, Finds Minimal Risk After Software Updates - 2026-04-07
