Core Viewpoint - The article discusses the journey of Li Auto in developing its autonomous driving technology, highlighting the challenges faced and the strategic decisions made to shift from relying on suppliers to self-research and development. Group 1: Historical Context and Initial Challenges - In 2020, Li Auto sold 32,624 units of the Li ONE, significantly exceeding internal expectations, which were initially set at 3,000 units for the first year [2][6] - Despite the success, the company faced immediate pressure from competitors like NIO and Xpeng, who were launching new products and advanced autonomous driving features [4][6] - The internal celebration of sales success contrasted sharply with the external pressures of market competition and technological gaps [6][7] Group 2: Decision to Pursue In-House Development - The decision to pursue in-house development of autonomous driving technology was driven by the realization of dependency on suppliers and the need for greater control over technology [9][19] - Li Auto's leadership recognized the necessity of self-research to remain competitive, especially after observing advancements made by competitors [22][23] - The initial budget for autonomous driving research was limited, forcing the team to be resourceful and strategic in their approach [9][10] Group 3: Development Process and Key Milestones - The "Fortress Project" was initiated to develop Li Auto's first fully self-researched advanced driver assistance system, with a tight deadline of less than 100 days [27][30] - The team faced significant challenges, including high turnover and the need to recruit quickly to meet project demands [32][30] - The successful delivery of the autonomous driving system before the launch of the 2021 Li ONE marked a significant achievement for the company [34][44] Group 4: Data-Driven Approach and Technological Advancements - The establishment of a data closed-loop system named "Poseidon" was crucial for enhancing the efficiency of the autonomous driving development process [39][40] - The data-driven approach allowed the team to rapidly iterate and improve the autonomous driving features based on real-world data [41][42] - By the end of 2021, Li Auto achieved a delivery volume of 90,491 units, a 177.4% increase year-on-year, largely attributed to the new self-researched driving system [43][44] Group 5: Ongoing Challenges and Future Plans - The company faced ongoing challenges in keeping pace with competitors in the rapidly evolving autonomous driving market, particularly in urban navigation capabilities [51][52] - Li Auto's strategic pivot towards end-to-end development and the initiation of new projects like the "Golden Apple Plan" reflect its commitment to innovation and competitiveness [58][59] - The article concludes with anticipation for future developments in Li Auto's autonomous driving technology, emphasizing the need for continuous adaptation and strategic foresight [60]