Humans need a driver’s test. Robot drivers mostly need a press release. On Sep 20, TechCrunch Mobility highlighted a federal attempt to change that: the National Highway Traffic Safety Administration (NHTSA) struck an agreement with SAE to launch ASCEND, a government-industry consortium meant to accelerate data-driven performance standards for automated driving systems (ADS).
The core problem is embarrassingly basic. There is still no standardized test for judging how well an autonomous system drives in the real world. Companies deploying robotaxis, autonomous trucks, and eventual personally owned AVs have largely answered “safe enough” for themselves—while states pile on inconsistent local rules.
Why existing car rules do not fit
Federal Motor Vehicle Safety Standards (FMVSS) already govern conventional vehicles. They were not written for designs like Tesla’s Cybercab or Zoox robotaxis that skip federally familiar equipment: pedals, steering wheels, side mirrors, rearview mirrors. Regulating those machines with yesterday’s checklist means either endless exemptions or forcing the agency into what Administrator Jonathan Morrison called “whack-a-mole” policymaking.
NHTSA says it wants a federal automated-vehicle framework that puts safety first without pretending it can predict every architecture that rolls out of a garage in 2028. That is a hard recipe: too vague and cities keep inventing their own rules; too rigid and the next vehicle form factor needs another exemption letter.
What ASCEND is supposed to deliver
Speaking at the AI Horizons Summit in Pittsburgh, Morrison said the consortium’s top priority is ADS competency—metrics and methods that can inform and accelerate a corresponding federal motor vehicle safety standard.
“By delivering an AV competency standard, we hope we’ll be able to address the burgeoning past patchwork of inconsistent state and local level requirements, and continue to encourage American innovation and promote the safe deployment and development of these vehicles on the road,” he said, per TechCrunch.
There is no public timeline yet for when ASCEND produces those standards. Industry will almost certainly try to shape the benchmarks; that is both the point of a consortium and the risk of one. Competency defined only by the companies being tested is how you get grade inflation with lidar.
The market context ASCEND walks into
TechCrunch’s same Mobility roundup underscored how commercial the AV race already is while the federal test is still being invented:
- May Mobility plans to go public via SPAC at a roughly $1.4 billion valuation, aiming to be the first U.S. pure-play public robotaxi company on an “asset-light,” partnership-first model.
- Waymo restarted San Antonio service five months after a flood incident, launched in Las Vegas, and said it is heading to Singapore for mapping and testing.
- Zoox’s 100-vehicle Las Vegas cap expires Sep 25, opening room to expand its custom robotaxi fleet.
- Lucid partnered with Europe’s Bolt to deploy at least 25,000 fully autonomous vehicles on an upcoming midsize EV architecture.
- Beep merged with Miami on-demand transit firm Freebee, consolidating another autonomy-plus-operations story.
Competency standards arriving after fleets already scale is the classic autonomy pattern: deploy first, define “safe enough” in public later. ASCEND is NHTSA admitting the pattern cannot continue forever if the agency wants federal leadership instead of fifty state science projects.
Geeknewz take
ASCEND will not magically end AV crashes or state fights. It is an admission that “trust us, our miles look good” is not a durable federal posture. If NHTSA can turn consortium work into a real ADS performance rule, the U.S. gets something closer to a license test for software drivers. If it stalls, the patchwork wins—and every city remains its own regulator with a different definition of safe enough.
Watch two things next: whether ASCEND publishes measurable competency criteria the public can understand, and whether public robotaxi operators treat those criteria as a product requirement or a lobbying target.
Source: TechCrunch Mobility — How do we know when an AV is safe enough? (Kirsten Korosec, Sep 20, 2026).