Tesla’s autonomous-driving ambitions are advancing within a rapidly changing regulatory, competitive, and infrastructure environment. The company’s Full Self-Driving (FSD) and robotaxi programs have moved beyond limited demonstrations toward broader, geofenced deployment, but their progress remains incremental. Technical shortcomings, unresolved hardware constraints, and fragmented approval regimes continue to limit the pace of unsupervised operation.
At the same time, Tesla’s Supercharger network remains a strategic asset as the industry transitions from CCS to NACS. That advantage is not unqualified: connector incompatibilities, uneven charging readiness, and expanding third-party competition are narrowing the gap. The central question for Tesla is therefore not whether it can demonstrate technological progress, but whether it can convert that progress into safe, scalable, and commercially durable systems. The proof is in the performance, not the promise.
Autonomous Driving: Progress Within Defined Boundaries
Tesla’s unsupervised robotaxi program, referred to as Cybercabs, now comprises approximately 21 vehicles operating across Austin, Dallas, and Houston 7,28. Operations are generally limited to the period from 6 a.m. to 2 a.m. and are suspended during rain 31. These restrictions illustrate the practical importance of the operational design domain: a system may function in a controlled geography and under favorable conditions without yet demonstrating general-purpose autonomous capability.
Tesla has also installed annual capacity at Giga Texas for more than 125,000 purpose-built Cybercabs 39, while prototype testing has expanded to multiple U.S. cities 27. This manufacturing preparation signals confidence in future scale, but production capacity is not the same as validated operational capability. Safety engineering is what happens between the edge cases—particularly the weather, road configurations, and unusual interactions that fall outside a carefully bounded demonstration.
On the software side, FSD v14 Lite has begun rolling out to vehicles equipped with Hardware 3 (HW3) 12,33. Early user reports remain mixed, citing phantom braking 18,35, lane ping-ponging 18,35, and unreliable parking behavior 18,36. Other users report improvements in turning, unprotected left turns, and reaction time 37. These accounts point to a system undergoing refinement rather than one that has eliminated the fundamental uncertainties of conditional automation.
The rollout also exposes a continuing tension between Tesla’s camera-only approach and the sensor redundancy favored by some regulators. New Jersey, for example, has proposed requiring radar and lidar for driverless commercial vehicles 4. If a system cannot guarantee an appropriate response to an edge case under degraded visibility or ambiguous roadway conditions, can it properly be treated as ready for unrestricted commercial operation? The answer must come from hazard analysis, validation suites, and demonstrated fault tolerance—not from the breadth of a marketing label.
International access to FSD Supervised has expanded into several European and Asia-Pacific countries 19, but approvals remain fragmented. France and Canada are still working through authorization processes 5,18. Hardware limitations, particularly in older HW3 vehicles, add another constraint 11. Customers also cannot transfer FSD purchases to newer models 34, complicating the transition from legacy hardware and adding friction to the customer experience.
Regulatory Exposure Is Becoming a Scaling Constraint
Tesla’s autonomous expansion is occurring under increasingly intensive regulatory scrutiny. NHTSA has opened a formal investigation into door safety following reports that owners were unable to enter their vehicles, in some cases with children inside 6. The agency has requested extensive internal documentation, including company communications and technical assessments 29. Although NHTSA has chosen to pursue rulemaking rather than a defect investigation in this matter 6,28, new federal motor vehicle safety standards could ultimately impose additional design or hardware requirements.
A separate NHTSA probe into unexpected braking, initiated in 2022, has accumulated 300 incident reports 21. Taken together, these proceedings demonstrate why certification cannot be treated as a one-time administrative hurdle. Standards are living documents, and safety compliance investigations are mechanisms for adapting them when field performance reveals failure modes that laboratory testing did not fully capture.
The regulatory environment is also divided across jurisdictions. Texas and Florida permit unattended operation 27, while British Columbia bans Level 3 autonomous vehicles and provides for potential jail time 30. Quebec, by contrast, permits Level 3 vehicles under existing law, allowing a software update to activate such features immediately 30. This asymmetry complicates national deployment strategies and increases the compliance burden for any manufacturer seeking to operate across borders.
For Tesla, the practical implication is clear: regulatory uncertainty can delay deployment even when the underlying software improves. Every marketed capability carries a corresponding duty of care. The appropriate response is not merely to seek faster approvals, but to establish evidence that aligns the system’s actual capabilities, operating boundaries, and failure responses with each jurisdiction’s requirements.
Superchargers: Strategic Advantage Amid a Connector Transition
Tesla’s charging network remains one of the company’s strongest ecosystem assets. The Supercharger network reports 99.95% uptime in North America 19, and V4 stalls are expanding with charging capability of up to 325 kW 23. Such reliability and higher-power equipment can reduce charging friction and support broader EV adoption.
Yet the industry-wide transition from CCS to NACS is creating compatibility bottlenecks. Only Gen 3 and newer Superchargers support non-Tesla EVs, and adapters remain necessary 25. Some vehicles, including the Chevrolet Bolt and Nissan Leaf, have port locations that are incompatible with the short cords at Gen 3 stations 24,25. In these cases, a nominally open network does not necessarily provide a seamless charging experience. The signal integrity of the system includes the connector, cable reach, vehicle port, payment process, and the charger’s operating reliability.
Third-party networks—including Ionna, EVgo, and Electrify America—are also expanding 22. Reliability and payment friction remain recurring weaknesses, however. Tesla’s advantage therefore persists, but it is being tested by a market in which access is broadening and the technical standards are converging. Certification should be a floor, not a ceiling; the same principle applies to charging performance, where nominal compatibility must be measured against actual usability.
Charging readiness remains uneven across regions. In Japan, only 5% of apartments offer EV charging 26, while timer-limited fast chargers remain common 32. Norway demonstrates the opposite dynamic: EVs account for 97% of new-car sales 10, showing how incentives and infrastructure can reinforce one another. However, the planned phase-out of VAT exemptions by 2028 20 points toward a normalization period in which adoption must increasingly stand on product economics and infrastructure quality rather than policy support alone.
Competition and Policy Are Reshaping the Market
Tesla faces growing pressure from both established automakers and Chinese entrants. BYD’s luxury Denza brand has launched the Z9 GT in Malaysia, with reported 1,500 kW flash charging and a 0–100 km/h time of 2.7 seconds 9,17. Xiaomi’s Sky Nomad is targeting lifestyle-oriented buyers with flexible interior layouts 16. BMW’s iX5 is advertised with a range of 435 miles and 10–80% charging in approximately 20 minutes 14. These products illustrate how competitors are narrowing the field through combinations of charging speed, range, performance, and design rather than relying on a single differentiator.
Battery chemistry may further alter the competitive balance. CATL and BYD are advancing sodium-ion battery technology 13, potentially reducing dependence on lithium and changing cost dynamics. If battery and charging technologies become more widely commoditized, Tesla’s long-term differentiation will depend increasingly on the integration of software, manufacturing, infrastructure, and safety performance.
The market is also being shaped by geopolitics and policy. The U.S. Department of Commerce denied Polestar authorization to sell vehicles in the United States because of concerns involving connected-vehicle software 15. Meanwhile, the expiration of the $7,500 federal EV tax credit 1,2,3,8 and a pending federal infrastructure ban on foreign inverters 38 introduce additional uncertainty for manufacturers and infrastructure providers.
Implications for Tesla
Tesla’s strategic position will be tested across three linked fronts. First, the company must demonstrate that FSD and robotaxi operations can scale without allowing unresolved edge cases to become field incidents. The current combination of limited geofences, weather restrictions, mixed user reports, and HW3 constraints suggests that widespread unsupervised operation remains a staged objective rather than an immediate outcome.
Second, Tesla must manage an increasingly complex regulatory map. NHTSA investigations, potential federal rulemaking, and divergent state and international requirements could impose compliance costs or hardware changes that affect deployment schedules. A robust certification pathway should therefore connect FMEA, hazard analysis, redundancy architecture, software validation, and post-deployment monitoring. These software boundaries are the new interlocking signals; they must be designed to fail safely when conditions exceed the system’s authority.
Third, Tesla must preserve the economic value of its Supercharger network while making access genuinely practical for a broader range of vehicles. High uptime and V4 expansion provide a foundation, but connector transitions, adapter requirements, cord limitations, and regional infrastructure gaps remain adoption bottlenecks. Monetization will depend on the network’s real-world convenience as much as on its installed footprint.
The immediate priorities are consequently practical: execute FSD deployments within clearly defined operating domains, resolve the technical and regulatory issues surfaced by NHTSA, and expand Supercharger access without sacrificing reliability. Tesla’s valuation and strategic moat will depend on whether it can turn these systems into dependable infrastructure. Lasting progress will come not from moving fastest, but from making every critical component—software, vehicle, charger, and rule—perform reliably when conditions are least forgiving.