Tesla CEO Elon Musk has long identified the ‘acid test’ for full autonomy as the ability to sleep through your commute. Yet recent data from Tesla’s fully driverless robotaxi fleet in Austin exposes a crash rate four times higher than average human drivers, spotlighting ongoing hurdles for unsupervised Full Self-Driving (FSD).
- Austin robotaxis crash once every 57,000 miles, four times more frequent than human drivers.
- 4 million Teslas with Hardware 3 cannot run unsupervised FSD despite software updates.
- Tesla retroactively added 'supervised' to FSD contracts amid ongoing legal challenges.
Market signal
Tesla’s deployment of fully driverless robotaxis in Austin represents one of the earliest real-world unsupervised driving tests by a major automaker. However, their reported crash frequency of one every 57,000 miles stands in stark contrast to advertised safety improvements. This disparity challenges Tesla’s broader narrative about the readiness of its Full Self-Driving technology and highlights technical and operational complexities still unresolved at scale.
Meanwhile, Tesla’s vast fleet—over 8 million vehicles running FSD-related software—operates only with mandatory human supervision, and roughly half the fleet lacks hardware capable of true unsupervised autonomy. This disconnect between marketing claims and hardware capability is further compounded by Tesla’s revised contract terms explicitly emphasizing the supervisory responsibilities of owners, signaling cautious market positioning around full autonomy timelines.
Operator impact
For operators contemplating autonomous fleet deployment, Tesla’s Austin robotaxi experience underscores significant safety and reliability challenges, particularly in unsupervised contexts. The increased incident rate may drive insurance costs, regulatory scrutiny, and operational risk management concerns, complicating the path to scalable deployment without active human oversight.
At the consumer level, the inability to enable unsupervised Full Self-Driving on millions of Teslas due to hardware constraints forces continued reliance on driver attentiveness. Buyers should anticipate ongoing updates but remain aware that the full self-driving experience Musk envisions remains a future prospect, not an immediate reality, with material implications for vehicle usage and feature expectations.
What to watch next
Tesla’s timeline for unsupervised Full Self-Driving availability has been pushed to late 2026 at the earliest, contingent on the rollout of the next software iteration, FSD v15. Industry watchers should closely monitor the performance of this update alongside any hardware refresh announcements that may address current computing limitations.
Regulatory oversight and legal actions, including a US class-action lawsuit around alleged false advertising on FSD capabilities, are likely to influence Tesla’s deployment strategy and market communication. Additionally, international safety bodies like the European Transport Safety Council pressing for precautionary measures may impact Tesla’s expansion and adoption in global markets.