Tesla tests new self-driving model, eyes September release

Tesla is developing a new version of its Full Self-Driving (FSD) software with a larger model size and improved video processing, CEO Elon Musk said on social media on August 6, 2025.

Musk indicated the upgraded FSD, which uses a larger AI model and enhanced video compression, could be ready for public release by the end of next month if testing is successful.

Tesla’s FSD is a partially automated driving system that still requires drivers to remain attentive and ready to intervene.

The feature, offered as an upgrade to Tesla’s existing Autopilot, is based on AI models that process input from the car’s cameras and sensors.

Tesla reported a 16% drop in automotive revenue for the second quarter, and has seen significant declines in European sales.

🔗 Source: CNBC

🧠 Food for thought

1️⃣ Tesla’s FSD timeline promises face historical credibility gap

Musk’s latest promise of FSD readiness “end of next month” follows a pattern of optimistic timelines that haven’t materialized.

In 2019, Tesla anticipated launching fully self-driving cars by Q2 of the following year, claiming the technology would allow driverless operation without human intervention1. That timeline missed by years, with the current system still requiring driver supervision.

Even earlier, Tesla removed its Full Self-Driving option from vehicle configurators in 2018, with Musk confirming it was to streamline the purchase process while promising the functionality would arrive through over-the-air updates once regulatory approval was obtained2.

The recurring pattern of aggressive FSD timelines suggests Tesla continues to prioritize ambitious public commitments over conservative engineering estimates, despite the technology’s persistent challenges in achieving true autonomous operation.

2️⃣ Chinese competitors rapidly close Tesla’s FSD advantage

Tesla’s “10X params” upgrade comes as Chinese automakers deploy competing self-driving systems that challenge its technological leadership.

Ji Yue’s PPA system uses a camera-based approach similar to Tesla’s FSD, though it combines this with some radar backup and high-definition maps for urban navigation3. While Ji Yue has faced sales struggles with only 2,300 units sold by April 2024, the technical capability demonstrates rapidly advancing Chinese autonomous driving development.

More significantly, Huawei’s Advanced Driver Assistance System integrates multiple sensors including LiDAR and radar, while BYD offers cost-effective ADAS systems that are gaining traction in budget EV markets4.