At the 18th Xuanyuan Automotive Bluebook Forum held from May 15th to 17th in Guangzhou under the theme "Turning Point," He Xiaopeng, Chairman and CEO of XPeng Group, discussed the impact of large models on the autonomous driving industry. He stated that end-to-end technology is relatively easier to develop, lowering the entry barrier, but it does not truly advance autonomous driving to driverless levels. "It merely creates a better L2 system, not driverless technology."
This presents a significant challenge to the industry: if the outcome is only an improved L2 system, what is its value to customers, and where does the competitive advantage lie?
He further pointed out that the key to determining whether a company has made substantial progress in AI lies in its data volume, training scale, and changes in computing power. The most painful aspect of this process is the massive financial investment required.
"Last year, XPeng, including our flying car projects, spent approximately 11 billion yuan on R&D, with a considerable portion allocated to the AI field," He Xiaopeng admitted. He acknowledged that while this substantial expenditure has yielded significant results, it was also a period of "immense anxiety." However, he emphasized that the most crucial takeaway is: failing to embrace change inevitably leads to obsolescence.
Comments