Tesla is preparing to introduce a significantly enhanced version of its Full Self-Driving (FSD) software, with CEO Elon Musk confirming that the upcoming model will feature a tenfold increase in parameters. The move marks Tesla’s latest step toward advancing autonomous driving technology, with a possible public release targeted for the end of next month, pending successful testing.
The new FSD version is currently being trained and tested, with a prototype already operating in Tesla’s Robotaxi platform in Austin, Texas. Although the vehicles do not require driver supervision, a Safety Monitor remains seated in the passenger’s side during operations.
“We are expecting to increase the parameter count to what we think can probably tenfold the parameter count,” Musk said during Tesla’s second-quarter earnings call in July. “This is a tricky thing to do because as you increase the parameter count, you get to choke on memory bandwidth.”
Musk reiterated the statement in a recent post on X (formerly Twitter), saying the next-generation FSD version could be released by the end of September if testing proceeds as planned.
The parameter increase is expected to boost the neural network’s learning capabilities and improve decision-making in real-world scenarios. However, Tesla must also manage potential trade-offs, particularly in terms of system efficiency and memory limitations. “But we currently think we can tenfold the parameter count from what people are currently experiencing,” Musk added during the earnings call.
Beyond performance, the upgrade could support Tesla’s push to expand FSD availability beyond the U.S. into markets such as Europe and Asia. While technical development continues, regulatory approvals remain a significant hurdle for broader deployment.
Tesla continues to collect vast amounts of real-world driving data, which it uses to refine the FSD system’s accuracy and safety. The company sees autonomous capability as key to its long-term strategy, including the rollout of a commercial robotaxi network.
