Tesla has achieved a significant milestone in its pursuit of autonomous driving, with drivers accumulating over 8 billion miles using the Full Self-Driving (Supervised) beta software. This achievement marks a crucial step in the company’s ongoing efforts to refine its advanced driver-assistance system and move closer to achieving unsupervised self-driving capabilities.
Cumulative Mileage and AI Development
The accumulation of these miles is strategically important for Tesla, serving a dual purpose. Firstly, it represents a substantial real-world usage benchmark for the Full Self-Driving (Supervised) system. Secondly, it directly contributes to the company’s ambitious goal of gathering 10 billion miles of training data.
Tesla announced on Wednesday, via its official X account, that its fleet had officially surpassed the 8 billion-mile mark. This comes just a few months after reaching the 7 billion cumulative mile milestone, which was reported on December 27, 2025.
Tesla owners have now driven >8 billion miles on FSD Supervised https://t.co/0d66ihRQTa pic.twitter.com/TXz9DqOQ8q
— Tesla (@Tesla) February 18, 2026
The pace at which Tesla is gathering these miles is accelerating, indicating increasing adoption and usage of the FSD (Supervised) system by its customer base.
The Road to Unsupervised Driving
CEO Elon Musk has previously stated that approximately 10 billion miles of training data would be the threshold required to enable unsupervised self-driving. In January, Musk highlighted the complexity of achieving this goal, noting, “Reality has a super long tail of complexity.”
Training data for Tesla’s AI models primarily consists of the accumulated real-world miles driven by its fleet. This data is essential for training and enhancing its end-to-end AI systems. The real-world driving data captures a wide array of unpredictable, rare, and complex scenarios that are difficult to replicate accurately and at scale solely through simulations.
Distinction Between FSD Usage and Training Data
It is important to distinguish between the total miles driven using the Full Self-Driving (Supervised) suite, which has now surpassed 8 billion, and the 10 billion miles of training data Musk has cited as necessary. While the miles driven on FSD (Supervised) significantly contribute to the training dataset, the 10 billion-mile figure represents an estimated cumulative real-world exposure needed to achieve human-level reliability for the autonomous system.
The FSD (Supervised) suite is recognized as one of the most advanced and data-rich semi-autonomous driving systems currently available to the public. The continuous collection of driving data is fundamental to Tesla’s strategy of iteratively improving its AI and advancing its autonomous driving capabilities, aiming to eventually eliminate the need for human supervision.