Source: Tesla

Tesla (NASDAQ:TSLA) is currently training a new iteration of its Full Self‑Driving (FSD) system that features approximately ten times the number of parameters compared to the existing model.

Elon Musk revealed the development in an August 6 post on X, noting that the update, if testing proceeds smoothly, could be released to the public by the end of next month.

Source: X @Elon Musk

This announcement marks a significant change from previous FSD updates. Earlier in the year, Tesla suggested a fourfold parameter increase, but the company appears to be skipping that interim step in favor of a more substantial jump.

The tenfold increase is intended to enhance the model’s ability to process visual inputs, aided by an improvement in video compression loss.

Tesla’s FSD suite has long been central to the company’s strategy of combining software and hardware innovation.

The firm’s Robotaxi service, launched in Austin earlier this year, already employs advanced versions of FSD. By training a model with such a significant scale-up, Tesla is effectively aiming to unify performance improvements from its commercial autonomous efforts with public vehicle capabilities.

Technically, the new model’s parameter expansion is expected to empower the network to capture more nuanced visual information and handle complex driving scenarios.

This architectural leap comes with challenges, as increased parameters can strain memory bandwidth and computational capacity. The company’s Dojo supercomputer infrastructure, designed specifically for large-scale neural network training, likely supports this process.

Although Tesla has not disclosed other technical details, the aim appears clear: a richer neural network that interprets real-world inputs more effectively. Past FSD versions focused on clearer lane-keeping and navigation.

Now, the added visual fidelity may enable capabilities such as pothole avoidance, improved edge-case handling, smoother traffic movement, and better parking guidance, as speculated by observers.

Existing FSD users, especially those on Hardware 4 platforms, have not seen significant gains in disengagement rates in recent months, based on community data.

The expectation is that the new model could reverse this trend and bring measurable improvements in real-world safety metrics.

This upgrade is being pursued during heightened scrutiny of Tesla’s autonomous systems. A recent legal verdict in Florida assigned $243 million in damages to Tesla in a fatal Autopilot crash, and new shareholder lawsuits challenge the company’s claims about Robotaxi and FSD performance.

Tesla did not provide pricing, delivery, or release timeline specifics beyond the general expectation of a public rollout "at the end of next month."