新能源车企激战智驾正酣之际，小鹏汽车(09868)传出智驾一号位换帅的消息。 近日，小鹏汽车自动驾驶中心发布内部信，李力耘不再担任自动驾驶中心负责人，世界基座模型负责人 刘先明接替该职位。 刘先明于2024年3月加入小鹏汽车，担任AI团队负责人。此前，他在2016年于伊利诺伊大学厄巴纳-香槟 分校电气和计算机工程专业博士毕业后，曾任职于Meta、通用汽车旗下的自动驾驶公司Cruise等，担任 机器学习与计算机视觉领域的研究工作。 在小鹏任职期间，刘先明推动了自动驾驶技术向AI大模型方向的演进。在今年6月的CVPR WAD上，他 发表了题为《通过大规模基础模型实现自动驾驶的规模化》的演讲，系统地介绍了小鹏汽车自研的业界 首个超大规模自动驾驶基座模型的历程和方法。 历经多次主将更换的小鹏汽车自动驾驶业务，也落地了不少进展。早在2022年9月，小鹏率先在广州开 启城市NGP(智能导航辅助驾驶)试点，高阶智能驾驶由此在城市场景量产落地。在此之后，NGP可以说 成为了小鹏自动驾驶的核心卖点，截至2024年初XNGP在全国243座城市全面开通，可用道路里程达56.9 万公里。 战略上，小鹏的自动驾驶业务也在深化，从202 ...

21世纪经济报道记者 吴佳楠 深圳报道 9月15日，粤港澳大湾区生成式人工智能安全发展联合实验室揭牌仪式在河套深港科技创新合作区举行。 联合实验室汇聚多方力量共同建设，参与单位包括粤港澳有关部门，腾讯、OPPO、小鹏汽车、深信服、视源、云蝶、云天励 飞、元象、深译、陆兮等AI应用龙头企业和AI专精特新企业，中山大学、广州大学、工信部电子五所等高校和科研院所。 联合实验室有关负责人向记者表示，联合实验室立足粤港澳大湾区科技创新高地，通过整合政府、企业、高校、科研机构及媒 体资源，构建"政产学研用"深度融合的创新生态，本质上是一种为适应人工智能时代而构建的"动态敏捷、多元协同"的创新治 理联合体。 聚焦三大领域发挥优势 当前随着人工智能技术向千行百业深度渗透，人工智能安全风险呈现跨领域、复合型、动态化特征，传统"事后监管"模式已难 以应对复杂挑战。 从技术层面看，生成式人工智能的随机性输出导致传统安全防护失效，传统规则引擎无法覆盖大模型的涌现风险，比如深度伪 造内容、隐蔽性数据泄露等；从生态层面看，生成式人工智能的对抗样本攻击、越狱攻击等新型威胁手段迭代加速，而多数市 场主体尤其是中小微企业的安全防御技术储备不足 ...

一手抓发展，一手抓安全，广东有两个大动作，透露了人工智能产业发展的最新风向。 就在今天，粤港澳大湾区生成式人工智能安全发展联合实验室在深圳正式揭牌。 这个联合实验室有多牛？看看它的"朋友圈"就知道了。 腾讯、小鹏、OPPO、视源股份（002841）等众多AI领域的头部企业，以及中山大学、广州大学、工信 部电子第五研究所等高校和科研院所，都参与其中。 广东、香港、澳门相关部门深度参与，汇聚多方力量共同建设。 如此高规格、跨区域、全方面的协作格局，信号很不寻常——这并不是一个普通的科研项目，而是吹响 了由广东主导、面向全球AI格局的"战略集结号"。 为什么要搞这么一个围绕人工智能安全发展的"超级联盟"呢？ 从顶层设计来看，发展与安全并重已经成为我们人工智能的发展基调，国家明确提出，要"完善生成式 人工智能发展和管理机制""建立人工智能安全监管制度"。 今年8月，国家印发《关于深入实施"人工智能+"行动的意见》这一重磅文件，为AI产业高质量发展按 下加速键。 而广东，作为改革开放前沿、经济第一大省，选择率先破局。联合实验室的真正使命，可以用几 个"最"来概括： 努力实现属地企业安全合规成本全国最低； 陆兮科技，推出 ...

Core Viewpoint - Xiaopeng Motors has achieved significant pre-order success with the new Xiaopeng P7, receiving 10,000 orders in just 7 minutes, indicating strong market demand and consumer interest [1][39]. Group 1: Product Launch and Market Performance - The new Xiaopeng P7's pre-order performance is compared to other models, with the Xiaomi SU7 achieving 10,000 orders in 4 minutes and the Xiaopeng MONA M03 in 52 minutes, showcasing the competitive landscape [3]. - Xiaopeng Motors reported a total delivery of over 100,000 units in Q2 this year, marking a historical high with a year-on-year growth of 241.6% [39]. - The company's revenue reached RMB 18.27 billion in Q2, also a record high, with a year-on-year increase of 125.3% [39]. Group 2: Technological Strategy and Development - Xiaopeng Motors has shifted from supporting laser radar to a firm commitment to a pure vision approach for autonomous driving, aligning with Elon Musk's recent endorsement of the same technology [8][10]. - The company claims its self-developed Turing chip has effective computing power equivalent to three NVIDIA Orin X chips, positioning it as a leader in the industry [23][30]. - Xiaopeng's self-developed chips are tailored for their products, allowing for higher efficiency and better performance compared to generic chips used by competitors [28]. Group 3: Future Outlook and Competitive Landscape - Xiaopeng Motors anticipates launching L4 autonomous vehicles by 2026, with plans for pilot robotaxi operations in select areas [34]. - The CEO of Xiaopeng Motors stated that the sales in the past year and a half equate to the total sales of the previous nine years, indicating a strong growth trajectory [39]. - The company emphasizes its focus on technology and aesthetics as key drivers for future development, while acknowledging the competitive challenges ahead [40][42].

Core Viewpoint - Xiaopeng Motors has achieved significant sales success with the new Xiaopeng P7, receiving 10,000 pre-orders in just 7 minutes, indicating strong market demand and consumer interest [2][56]. Group 1: Sales Performance - The new Xiaopeng P7 received 10,000 pre-orders in 7 minutes, showcasing its popularity compared to other models like the Xiaomi SU7 and Xiaopeng MONA M03, which had longer pre-order times [2][4]. - Xiaopeng Motors reported that its sales in the past year and a half are equivalent to the total sales of the previous nine years, highlighting a remarkable growth trajectory [56]. Group 2: Technological Advancements - Xiaopeng Motors has shifted from supporting laser radar to a pure vision approach for autonomous driving, aligning with industry trends and competitors like Tesla [13][26]. - The company claims its self-developed Turing chip has an effective computing power equivalent to three NVIDIA Orin X chips, significantly enhancing its autonomous driving capabilities [38][41]. - The advancements in computing power are expected to improve the performance of pure vision systems, allowing them to handle complex driving scenarios better than before [33][34]. Group 3: Future Outlook - Xiaopeng Motors aims to launch L4 autonomous vehicles by 2026 and is considering partnerships for robotaxi operations, indicating a strategic approach to future mobility solutions [49][54]. - The company anticipates a significant update in its autonomous driving capabilities by the end of this year, claiming it will outperform competitors by a factor of ten [49][56]. - Despite the positive outlook, the CEO acknowledges that the automotive industry remains competitive, with no company guaranteed success [60][62].

Core Viewpoint - Xiaopeng Motors reported better-than-expected automotive gross margins in Q2 2025, with revenue of 18.27 billion RMB, a 15.6% increase quarter-on-quarter, and vehicle sales of 103,000 units, a 9.8% increase quarter-on-quarter [1] Group 1: Financial Performance - Q2 2025 automotive gross margin reached 14.3%, a significant improvement of 3.9 percentage points quarter-on-quarter, exceeding market expectations [1] - R&D and sales expenses increased by 11.4% quarter-on-quarter, but were lower than revenue growth, leading to a slight decrease in expense ratio [1] - The company reported a net loss of 480 million RMB and a non-GAAP net loss of 380 million RMB, with losses narrowing further quarter-on-quarter [1] Group 2: Future Guidance - For Q3 2025, the company expects revenue between 19.6 billion and 21 billion RMB, indicating a quarter-on-quarter growth of approximately 11.9% [2] - Expected delivery volume for Q3 2025 is between 113,000 and 118,000 units, reflecting a quarter-on-quarter growth of about 11.1% [2] - The anticipated average selling price (ASP) for Q3 2025 is around 170,000 RMB, a slight increase of 6,000 RMB quarter-on-quarter, suggesting potential improvements in vehicle mix and gross margin [2] Group 3: Strategic Developments - The company maintains its guidance for profitability in Q4 2025, with expectations for continued upward trends in ASP and gross margin [2] - The new generation P7 has been unveiled and is set for delivery in Q3 2025 to Q4 2025, while the "Kunpeng Super Electric System" is expected to enter mass production in Q4 2025 [2] - The self-developed Turing chip is set to be deployed in Q3 2025, with plans to extend its application to robotics and other product lines by 2026 [2] Group 4: Investment Outlook - The company is viewed as a high-certainty investment among new energy vehicle manufacturers, with expectations for continued sales growth driven by new and updated models [3] - The potential for quarterly profitability is anticipated due to the proliferation of intelligent driving and new vehicle launches [3] - The company's strategic positioning in artificial intelligence, low-altitude economy, RoboTaxi, and humanoid robots is expected to positively impact stock prices and elevate valuation [3]

Core Viewpoint - The article discusses the recent challenges faced by Nvidia in the Chinese market following the U.S. lifting the ban on its H20 AI chip sales to China, highlighting concerns over potential security risks associated with backdoor systems in chips [1][12]. Group 1: Nvidia and H20 Chip - Nvidia's H20 chip, designed for the Chinese market, is now facing scrutiny from China's National Cyberspace Administration due to concerns about security vulnerabilities and potential backdoors [1][12]. - The H20 chip's performance is estimated to be about 70% of Nvidia's H100 chip, making it the most powerful AI chip Nvidia is allowed to sell in China [12]. - Despite receiving 300,000 orders, the scrutiny from the Chinese government poses significant challenges for Nvidia's sales strategy in the region [12][14]. Group 2: Backdoor Systems and Security Risks - Backdoor systems in chips can allow unauthorized access, posing severe security threats, especially in critical applications like military and finance [2][4]. - The definition of "backdoor" is contentious, with some features being misidentified as malicious due to their potential misuse [4][5]. - Experts emphasize that distinguishing between design flaws and intentional backdoors requires precise technical analysis [5][6]. Group 3: Geopolitical Context - The scrutiny of Nvidia's H20 chip reflects broader geopolitical tensions between the U.S. and China, particularly in the tech sector [12][15]. - China's emphasis on technological self-sufficiency and reducing reliance on Western technology is becoming increasingly pronounced [12][15]. - The incident illustrates the disconnect between technological trust and geopolitical trust, amplifying concerns over security in international tech collaborations [15].

Core Viewpoint - The delay of NVIDIA's Thor chip has significantly impacted domestic automakers, particularly Xiaopeng and Li Auto, who were relying on its capabilities for their models [1][5][9]. Group 1: NVIDIA Thor Chip Issues - NVIDIA's Thor chip, initially promised for mass production by the end of 2024, has faced design issues leading to low yield rates, with CEO Jensen Huang admitting the problem lies with NVIDIA [1][5]. - The promised single-chip performance of 2000 TOPS has been downgraded to 700 TOPS, with actual tests showing performance around 500 TOPS, which does not provide a competitive advantage over existing solutions [3][5]. - The delay in Thor's production has forced companies like Xiaopeng to switch to alternative solutions, such as dual Orin X chips, to meet their production timelines [5][7]. Group 2: Impact on Domestic Automakers - Xiaopeng's G7 model had to switch to a self-developed Turing chip due to the Thor chip's repeated delays, with only the Ultra version utilizing the Turing chip [7][8]. - Li Auto's VLA model, which requires the Thor chip's processing power for advanced road recognition, is also significantly affected, as it cannot be deployed without it [9][11]. - Both companies are now looking towards self-developed chips as a more reliable solution, with Li Auto's "Schumacher" chip expected to be ready by Q1 2026 [11][20]. Group 3: Shift Towards Self-Developed Chips - The trend towards self-developed chips is gaining momentum among domestic automakers, with NIO having started its chip development as early as 2021, resulting in the release of the Tianji NX9031 chip with 1000 TOPS performance [17][19]. - Xiaopeng's Turing chip, with a performance of 750 TOPS, is also in development, although it has not yet been fully optimized for autonomous driving [19]. - Huawei's Ascend 910 B chip, designed for L3 level driving assistance, is another example of the shift towards self-reliance in chip development [20][23]. Group 4: Industry Implications - The delays caused by NVIDIA's Thor chip have inadvertently provided an opportunity for domestic competitors to catch up in the autonomous driving chip market [30][32]. - The necessity for self-developed chips is emphasized as a means to enhance vertical integration and better manage chip performance and deployment [30][32]. - The long-term accumulation of technology in chip design and manufacturing is crucial for companies to avoid dependency on external suppliers like NVIDIA [32].

Core Viewpoint - The article discusses the competitive landscape of the autonomous driving industry, focusing on NVIDIA's challenges in maintaining its market position against emerging Chinese companies and the shift towards self-developed chips by major automakers [5][15][50]. Group 1: NVIDIA's Market Position - NVIDIA's market capitalization has reached $4 trillion, making it the world's most valuable company, but it faces increasing competition from Chinese automakers who are trying to reduce reliance on NVIDIA's technology [5][15]. - General Motors' executives have expressed concerns about NVIDIA's autonomous driving solutions, indicating potential issues in their collaboration [7][8]. - Other automakers, such as Mercedes-Benz, have also reported that NVIDIA's autonomous driving performance is lagging behind that of Chinese startups like Momenta [10][11]. Group 2: Challenges in Chip Delivery - NVIDIA's latest Thor chip has faced multiple delays, impacting key clients like Li Auto, which has resulted in significant sales losses estimated at around 6 billion yuan due to postponed vehicle launches [18][19]. - The delays in chip delivery have prompted companies like Xiaopeng to pivot towards self-developed chips, as they can no longer rely on NVIDIA's timelines [20][24]. - The challenges faced by NVIDIA in delivering the Thor chip are attributed to design flaws and the complexity of automotive-grade chip production, which differs from consumer electronics [34][42][46]. Group 3: Shift Towards Self-Developed Chips - Major Chinese automakers are increasingly investing in self-developed chips to reduce costs and enhance compatibility with their AI technologies, with companies like NIO and Xiaopeng already making significant progress [25][35][37]. - The self-development of chips is seen as a strategic necessity for automakers to maintain competitiveness in the rapidly evolving autonomous driving market [38][39]. - The article highlights that the development of self-developed chips is a long-term commitment, with significant investments and risks involved, but it is becoming essential due to supply chain uncertainties [26][27][30]. Group 4: Competitive Landscape - The competition in the autonomous driving software space is intensifying, with Chinese companies like Momenta and Qingtou Zhihang rapidly advancing their technologies, often outpacing NVIDIA's offerings [51][53]. - NVIDIA's corporate culture and operational structure may hinder its ability to adapt quickly to the demands of the automotive industry, contrasting with the agile approaches of Chinese startups [52][54]. - The article suggests that the future of autonomous driving will likely see a shift towards more localized solutions, with Chinese companies capturing a larger share of the market as they innovate faster and align more closely with automotive needs [55].

Core Viewpoint - The article discusses the challenges faced by NVIDIA in the automotive sector, particularly in the context of its partnerships with major car manufacturers and the increasing competition from Chinese companies developing their own chips and software solutions [3][5][18]. Group 1: NVIDIA's Automotive Business Challenges - NVIDIA's automotive business, while significant, accounts for less than 2% of its total revenue of $130.5 billion, indicating that it is a relatively small segment for the company [11][58]. - The collaboration between NVIDIA and General Motors has faced internal criticism, with GM executives describing NVIDIA's autonomous driving solutions as "very scary" [5][6]. - Other automakers, such as Mercedes-Benz, have also expressed dissatisfaction with NVIDIA's performance, leading to a shift towards competitors like Momenta for autonomous driving solutions [9][11]. Group 2: Competition from Chinese Companies - Chinese automakers are increasingly developing their own AI chips, with companies like NIO and Xpeng already delivering their self-developed chips, posing a significant threat to NVIDIA's market share [19][30]. - The article highlights that the delay in NVIDIA's Thor chip delivery has prompted companies like Xpeng to pivot towards their self-developed chips, indicating a loss of confidence in NVIDIA's ability to meet delivery timelines [24][25]. - The competitive landscape is shifting, with Chinese companies rapidly advancing in autonomous driving software and hardware, making it difficult for NVIDIA to maintain its previous dominance [66][68]. Group 3: Implications of Chip Development - The development of self-research chips by automakers is seen as a strategic necessity, driven by the need for cost reduction and better integration with AI capabilities [45][49]. - The article notes that the challenges faced by NVIDIA in delivering the Thor chip have inadvertently accelerated the self-development of chips among leading Chinese automakers [31][30]. - The long development cycle for automotive chips, which can take up to four years, contrasts sharply with the faster-paced software development cycles seen in the industry [33][50]. Group 4: Cultural and Operational Differences - NVIDIA's corporate culture, which emphasizes long-term technological advancements, may not align with the immediate delivery needs of automotive clients, leading to operational friction [51][62]. - The article points out that NVIDIA's team in China lacks decision-making power compared to its larger U.S. team, which may hinder its responsiveness to local market demands [65]. - The disparity in urgency and operational focus between NVIDIA and its automotive partners has created a gap that competitors are eager to exploit [67][68].