> ## Content Index > Fetch the complete content index at: https://bytevyte.com/llms.txt > Use this file to discover other available public pages before exploring further. # NHTSA Drafts Outcome-Based Driverless Vehicle Safety Standards - URL: https://bytevyte.com/nhtsa-drafts-outcome-based-driverless-vehicle-safety-standards/ - Published: 2026-09-22T20:42:20.000Z - Updated: 2026-09-22T20:42:20.000Z - Description: NHTSA is drafting federal driverless vehicle safety standards built on safety outcomes, backed by a $5M SAE consortium, as Tesla's Cybercab review escalates. - Author: Bytevyte Editorial - Tags: ai-beats **NHTSA** is drafting a new class of federal performance standards for automated vehicles. The agency says the existing Federal Motor Vehicle Safety Standards were written around cars with a human driver and do not translate cleanly to machines built without one. The plan would put driverless vehicle safety standards on outcome-based footing, defining what a vehicle must achieve instead of which components it must carry. The agency's reasoning is narrow and technical. FMVSS rules assume a conventional passenger-car layout, which makes them a poor fit for wheel- and pedal-free designs such as **Tesla**'s Cybercab or **Zoox**'s purpose-built robotaxi. Robotaxi operators, autonomous truck developers and teams working on personal autonomous vehicles have relied on real-world driving mileage to show their systems are safe. NHTSA now says that evidence is not sufficient by itself. Enforcement pressure is running in parallel. The agency escalated its Cybercab inquiry to a formal Special Order on September 17, 2026, requiring sworn answers from Tesla by September 30\. A separate Audit Query examines whether the vehicle's self-certification against human-driver standards holds up. ## The Case for New Driverless Vehicle Safety Standards The gap is not new. NHTSA finalized its first occupant protection standards for vehicles without driving controls in March 2022\. That step allowed cars with no steering wheel or pedals to be certified for sale. Those rules did not establish how well an automated driving system must perform the driving task itself. That omission shapes how exemptions work today. When NHTSA cleared Zoox this summer to operate up to 5,000 robotaxis for two years, the agency stated that no performance requirements exist for an automated driving system replacing a human driver. It also said it has no authority to demand a higher safety bar while deciding an exemption. The statutory test asks only whether an exempted vehicle is at least as safe as a compliant one. Zoox's route to that decision shows the friction. The Amazon subsidiary tried to self-certify in 2022 and federal regulators did not accept the filing. It applied formally for special approval in the fall of 2025 and waited months for a ruling. Tesla sits earlier in the same process and faces more resistance. Its Texas robotaxi operation has grown to 67 vehicles after 28 additions. Self-certification is the mechanism that makes the current system workable and awkward at the same time. Manufacturers attest that their vehicles meet applicable FMVSS rules, and NHTSA checks those attestations after the fact through audits. That model strains when the rule being certified assumes a driver who no longer exists. ## Data, Consortiums and the Missing Yardstick The agency's answer to the missing evidence base is a consortium. NHTSA signed a cooperative agreement with **SAE International** to run **A2SCEND**, a three-year, $5 million program whose data is intended to inform the first federal performance standard for automated vehicles. According to Administrator **Jonathan Morrison**, the work is meant to speed up data-driven AV performance standards without letting regulation fall behind deployment. NHTSA is separately preparing an advance notice of proposed rulemaking on what it terms "ADS competency." That is the information-gathering phase that precedes a draft rule. It does not yet define what a compliant automated driving system must do, which leaves operators, insurers and legal teams planning against a benchmark that does not exist. An advance notice of proposed rulemaking is the earliest step in federal rulemaking and carries no legal force. It signals intent and opens a comment window; a proposed rule, then a final rule, would follow. NHTSA has not published a timetable for either, which makes the agency's own schedule one of the unresolved variables in the program. The three-year term of A2SCEND sets a rough clock. If the consortium's data feeds a proposed rule and the rulemaking process runs its full course, a binding federal performance standard for automated driving systems is unlikely to be in force before the end of the decade. That keeps the exemption path as the main route to market in the interim. Nothing in the effort suspends current enforcement. Defect investigations, recall authority and crash-reporting duties remain in force. The Department of Transportation's AV 5.0 national strategy sets out a path toward wider deployment through 2030. | Date | Development | Status | | ------------------ | ----------------------------------------------------------------------------------- | -------- | | March 2022 | First occupant protection standards for vehicles without driving controls finalized | In force | | 2022 | Zoox self-certification attempt declined by federal regulators | Rejected | | Fall 2025 | Zoox files for special approval of its purpose-built robotaxi | Resolved | | Summer 2026 | Zoox cleared for up to 5,000 robotaxis over two years | Active | | September 17, 2026 | NHTSA Special Order requires sworn Tesla responses by September 30 | Pending | | September 2026 | A2SCEND consortium with SAE International, three years, $5 million | Underway | ## What Changes for Operators and Investors Outcome-based rules move the burden of proof. Mileage totals become a supporting exhibit rather than the argument, and the questions that decide approval shift toward test methodology, scenario coverage and third-party validation. Companies that can produce auditable safety evidence gain an advantage over those that cannot. The cost of assembling that evidence becomes a line item in every deployment budget. How the standard gets written matters as much as whether it gets written. A rule built around measurable outcomes, such as crash and injury rates per mile inside defined operating conditions, gives manufacturers latitude in architecture. A rule that reintroduces design requirements through testing protocols would recreate the problem NHTSA says it is trying to solve. NHTSA has positioned the framework as a way to keep safety at the foundation without slowing deployment. It draws a line between specifying components and specifying results. Whether that line holds will depend on the testing protocols the agency writes. The exemption route carries its own limits for now. Because the agency cannot impose a performance floor in exemption decisions, each approval is case-specific, and a favorable ruling for one vehicle design says little about the next. Zoox's two-year, 5,000-vehicle allowance and Tesla's pending certification review are separate proceedings with separate outcomes. Automated trucking has the most to gain from a clearer benchmark. Long-haul operators run fixed corridors and could accumulate the scenario data a performance standard would reward. They also carry the largest liability exposure per incident, which makes an undefined safety bar expensive to insure. Tesla's Texas robotaxi fleet, at 67 vehicles, sits at a scale where a certification decision would reset its deployment timeline without producing meaningful near-term revenue either way. That asymmetry runs across the industry, where the compliance decision sets the schedule more often than the technology does. Geography compounds the problem. Waymo's expansion in Japan runs into transport law that places liability with the operating partner, Nihon Kotsu, rather than with the technology provider. A single federal driverless vehicle safety standards regime would not resolve that. It would give manufacturers one domestic target to design against, rather than a sequence of case-by-case exemption decisions. Adjacent work shows the agency is comfortable with performance criteria. NHTSA's first pedestrian safety test results for 2027 models found 102 vehicles from 10 manufacturers meeting new performance thresholds. That testing regime grades outcomes rather than component design. The automated vehicle effort applies the same logic to the driving task. ## Why this matters For operators and their backers, the practical question has shifted from whether driverless vehicle safety standards will change to which evidence will count when they do. Until a performance rule is final, deployment in the United States still runs through exemptions that the agency cannot judge against a safety floor. Approval timelines and liability exposure stay tied to individual proceedings rather than a settled benchmark. A2SCEND is meant to close that gap, and its three-year term sets the earliest point at which an answer could arrive. *AI-generated image.* ## Related Articles - [NHTSA Opens Tesla Cybercab Safety Probe Into How 1,000 Driverless Robotaxis Were Certified](https://bytevyte.com/nhtsa-opens-tesla-cybercab-safety-probe-into-how-1-000-driverless-robotaxis-were-certified/) - [Tesla's NHTSA Cybercab Certification Faces a Sworn Deadline of September 30](https://bytevyte.com/teslas-nhtsa-cybercab-certification-faces-a-sworn-deadline-of-september-30/) - [Federal Tesla FSD Probe Widens to 3 Million Vehicles After 49ers Coach Crash](https://bytevyte.com/federal-tesla-fsd-probe-widens-to-3-million-vehicles-after-49ers-coach-crash-2/) ✔Human Verified --- *Researched and cross-referenced against primary sources by the Bytevyte editorial team. This article was generated with the assistance of artificial intelligence and reviewed by the Bytevyte editorial team.*