On January 30, Tesla released its fourth-quarter and full-year financial results: Q4 revenue reached $25.71 billion (approximately ¥186.59 billion), marking a 2% year-on-year growth.
For the full year, Tesla’s total revenue in 2024 amounted to $97.7 billion (approximately ¥709.82 billion), and the company delivered 1,789,226 vehicles.
During an interview following the earnings report, Elon Musk addressed the question, “What is the biggest challenge for bringing FSD (Full Self-Driving) to China?” He responded: “Bus lanes.”
“They are very complex. At certain times of the day, they are open, while at other times, they are restricted. If you accidentally enter a bus lane at the wrong time, you’ll immediately receive a fine. In any case, we will solve this issue.”
Musk also mentioned that FSD is possible in every market, but regulatory limitations prevent it from being fully realized this year. In the U.S., the unsupervised version of FSD will roll out this year, with nationwide deployment expected next year. By the end of next year, many countries around the world will also see the deployment of the unsupervised version of FSD.
Comparison of Autonomous Driving System Architectures of Leading Companies
In 2023, Tesla began formal training of its end-to-end model and released the FSD Beta V12 testing version in 2024. In comparison, Chinese tech giant Huawei employs a BEV (Bird’s Eye View) + Transformer + GOD (General Object Detection) network architecture, further enhanced by road topology reasoning networks, such as Graph Neural Networks (GNN). This approach has demonstrated excellent performance in practical applications while reducing the reliance on high-definition maps for autonomous driving. Xiaomi’s autonomous driving system utilizes a BEV + Transformer + Occupancy Network architecture for its underlying algorithms, fully integrating large model technology.
