广告
加载中

AI从L9 Livis开始 理想进入具身智能时代

李玉鹏 2026-05-20 09:52
李玉鹏 2026/05/20 09:52

邦小白快读

EN
全文速览

本文核心介绍了理想汽车通过全新L9 Livis落地具身智能战略,从传统车企向掌握物理世界智能化底层能力的科技企业转型的全过程,核心干货信息如下:

1. 核心战略路线:李想提出具身智能分上下半场,上半场是自动驾驶汽车,分为三个阶段:2018-2023年L2辅助驾驶,2023-2028年L3自动驾驶,2028-2033年L4无人驾驶;下半场是通用人形机器人,也分三个阶段推进,上半场自动驾驶积累的底层技术可直接用于下半场机器人研发,二者用户重合度较高。

2. 企业转型动作:理想重构研发体系,打破传统软硬件团队壁垒,把AI工具嵌入全业务流程,智能辅助驾驶模型迭代周期从两周缩短至一天,同时重新定义AI时代人才标准,挖掘能利用AI改造业务的核心人才。

3. 技术布局策略:理想坚持全栈自研芯片、模型、操作系统等核心能力,提前布局未来十年的竞争壁垒,应对当前智能电动车行业的同质化竞争。

本文为新能源汽车品牌布局下一代竞争力提供了清晰的参考,核心干货如下:

1. 消费与产业趋势判断:当前智能电动车行业在续航、配置、补能等领域已经趋同,接下来竞争将从单车产品力升级转向底层智能化系统能力竞争,具身智能是汽车产业与科技产业融合的下一个方向,汽车会成为AI进入物理世界的核心入口。

2. 产品研发方向:可参考理想的路线,依托汽车成熟产业基础积累感知、算力、模型、控制等底层能力,未来延伸到通用人形机器人赛道,二者底层能力通用,用户群体重合度高,可提前布局打开第二增长曲线。

3. 组织与技术策略:要围绕具身智能战略重构研发组织,打破软硬件团队壁垒,将AI嵌入全业务流程提升效率;坚持全栈自研整套核心系统,形成技术代差和产品差异化,前置建设长期竞争壁垒。

4. 投入参考:理想2025年研发投入113亿元,AI占比五成,2026年预计投入120亿元,AI仍占一半,可为品牌研发投入规划提供参考。

本文为新能源汽车领域的卖家梳理了行业新趋势、机会与风险,核心干货如下:

1. 行业变化与增长机会:当前智能电动车赛道已经进入同质化阶段,接下来竞争焦点转向底层智能化能力升级,具身智能车型、高阶自动驾驶会成为接下来高端市场的核心卖点,用户对新技术升级的接受度和付费意愿较高,卖家可提前布局相关车型的推广,抓住技术转型带来的销量增长。

2. 风险提示:头部品牌已经提前十年布局底层技术壁垒,行业技术迭代速度会明显加快,如果不能跟上转型节奏,无法给用户传递清晰的智能化价值,很容易在竞争中被拉开差距,需要及时调整产品推广重心。

3. 可学习的经验:理想一直以来都擅长把复杂技术包装进具体的家庭用户场景,精准击中用户需求,卖家也可以借鉴这种思路,把抽象的具身智能技术转化为用户可感知的场景体验,提升转化效率。

4. 需求变化:用户对汽车的需求已经从空间、配置升级转向智能化体验升级,需要及时调整选品和推荐逻辑,匹配用户新需求。

本文为汽车生产制造工厂指明了智能化转型方向和新商业机会,核心干货如下:

1. 产品生产设计需求变化:具身智能时代的汽车对生产制造提出了更高要求,需要更高精度的线控底盘、执行机构,更稳定的硬件整合能力,对生产的安全性、一致性要求远高于传统车型,工厂需要尽快调整生产标准,适配新的生产要求。

2. 数字化转型启示:理想已经用AI重构研发流程,将AI嵌入全业务链条,大幅缩短了产品迭代周期,工厂也可以参考这个思路,把AI引入生产、管理、研发全流程,解决跨部门协同效率低的问题,提升整体生产研发效率。

3. 新商业机会:随着具身智能产业发展,核心零部件比如传感器、线控系统、算力硬件的需求会持续增长,长期来看通用人形机器人的生产制造需求也会逐步释放,工厂可提前布局相关产能,积累智能化生产经验,开拓新的业务增长点。

本文梳理了车企转型具身智能过程中的痛点和行业趋势,给面向汽车产业的服务商指明了方向,核心干货如下:

1. 行业发展趋势:新能源汽车产业已经从传统的供应链整合阶段转向全栈自研底层技术的新阶段,越来越多车企开启向具身智能企业的转型,相关技术服务和解决方案的需求会持续快速增长,赛道空间广阔。

2. 客户核心痛点:车企转型过程中,面临传统研发体系壁垒打破难、AI落地业务流程难、跨部门协同效率低、AI适配人才不足等核心痛点,传统的单点服务模式已经无法满足车企的整体转型需求。

3. 解决方案开发方向:服务商可围绕AI研发工具、组织转型咨询、AI人才培养、模型训练基础设施等领域开发整体解决方案,帮助车企更快打通软硬件研发壁垒,缩短研发迭代周期,适配全栈自研的转型需求。

4. 技术布局方向:感知模型、芯片模型协同设计、整车操作系统、线控系统整合等领域会持续产生新的服务需求,提前布局相关技术就能抢占市场先机。

本文为汽车相关平台商梳理了车企转型带来的新需求、运营调整方向和风险,核心干货如下:

1. 行业新需求:当前车企都在推进具身智能转型,推出新一代智能化车型,需要平台匹配新的技术展示、内容传播、用户触达方式,用户对自动驾驶、具身智能相关内容的关注度持续提升,平台需要及时调整运营逻辑适配新需求。

2. 招商与增长方向:头部车企已经提前布局下一代技术,全新的具身智能车型会陆续面市,平台可提前对接转型头部车企,争取新车型首发、优先推广等权益,吸引对智能化感兴趣的高端用户,提升平台整体流量和交易额。

3. 运营管理调整:用户对智能汽车的关注点已经从价格、配置转向底层智能化体验,平台可调整内容板块、车型评价体系,增加技术解读、智能化体验评测相关内容,更好匹配用户信息需求。

4. 风险规避:具身智能目前仍处于发展早期,部分概念尚未大规模落地,平台要避免过度炒作概念,引导用户关注真实产品体验,防范概念泡沫带来的用户信任危机。

本文展现了汽车与AI融合的最新产业动向,对产业研究有较高参考价值,核心干货如下:

1. 全新产业动向:当前智能电动车行业陷入同质化迷茫,理想提出具身智能的全新赛道方向,明确了汽车作为AI进入物理世界第一入口的定位,梳理了从自动驾驶到通用人形机器人的完整产业路线图,打通了两个赛道的底层能力逻辑,开辟了汽车产业新的增长叙事。

2. 新商业模式与组织模式探索:理想打破了传统车企依赖供应商整合的模式,探索出全栈自研底层技术的新路径,通过前置高投入研发建立长期竞争壁垒;同时探索出用AI重构组织流程的新模式,重新定义AI时代的人才标准和研发体系,为科技制造企业转型提供了新样本。

3. 值得深入研究的新问题:全栈自研高投入模式的商业回报逻辑、具身智能落地过程中成本、可靠性、安全性等门槛的突破路径、AI如何改造传统制造企业的组织体系,都是值得深入研究的新产业问题,为研究提供了新方向。

返回默认

声明:快读内容全程由AI生成,请注意甄别信息。如您发现问题,请发送邮件至 run@ebrun.com 。

我是 品牌商 卖家 工厂 服务商 平台商 研究者 帮我再读一遍。

Quick Summary

This article details Li Auto's transformation from a traditional automaker to a tech company with core embodied intelligence capabilities for the physical world, centered on the rollout of its embodied intelligence strategy via the new L9 Livis. Key takeaways are as follows:

1. Core strategic roadmap: Li Xiang, founder of Li Auto, divides embodied intelligence into two phases. The first phase focuses on autonomous driving (AD) and unfolds in three stages: Level 2 ADAS from 2018 to 2023, Level 3 autonomous driving from 2023 to 2028, and Level 4 fully driverless vehicles from 2028 to 2033. The second phase will target general-purpose humanoid robots, also advancing in three stages. Core technologies accumulated in the AD phase can be directly transferred to robot development, and the two businesses share largely overlapping user bases.

2. Corporate transformation initiatives: Li Auto has restructured its R&D system to break down traditional silos between software and hardware teams, embedding AI tools across all business processes. This has cut the iteration cycle for intelligent AD models from two weeks to just one day. The company has also redefined talent requirements for the AI era, prioritizing candidates who can leverage AI to reshape business operations.

3. Technology positioning strategy: Li Auto remains committed to full-stack in-house development of core capabilities including chips, models and operating systems, building competitive moats for the next decade to counter widespread homogenization in the current smart EV industry.

This article provides clear strategic references for new energy vehicle brands looking to build next-generation competitiveness. Key insights are as follows:

1. Consumer and industry trend outlook: The smart EV industry has already converged in range, features, and charging infrastructure. Future competition will shift from individual product performance to competition in underlying intelligent system capabilities. Embodied intelligence is the next direction for convergence between the automotive and tech industries, positioning cars as the primary entry point for AI to access the physical world.

2. Product R&D direction: Brands can follow Li Auto's playbook: accumulate core capabilities in perception, computing power, model development and control based on the mature automotive industry foundation, then expand into the general humanoid robot track in the future. Core capabilities are transferable between the two fields, and they share highly overlapping user groups, enabling early布局 to unlock a second growth curve.

3. Organizational and technical strategy: Brands should restructure R&D organizations around an embodied intelligence strategy, break down software-hardware silos, and embed AI across all business processes to boost efficiency. They should also pursue full-stack in-house development of core systems to create technological gaps and product differentiation, building long-term competitive moats in advance.

4. R&D spending benchmark: Li Auto plans to spend RMB 11.3 billion on R&D in 2025, with 50% allocated to AI; in 2026, R&D spending is projected to rise to RMB 12 billion, with AI again accounting for half. This provides a useful reference for brands planning their own R&D investment roadmaps.

This article outlines new industry trends, opportunities and risks for new energy vehicle sellers. Key takeaways are as follows:

1. Industry shifts and growth opportunities: The smart EV market has entered a phase of widespread homogenization, and competition will next shift toward upgrades to underlying intelligent capabilities. Embodied intelligence-enabled vehicles and high-level autonomous driving will become core selling points for the high-end market going forward. Consumers show high acceptance and willingness to pay for new technology upgrades, so sellers can start preparing to market related models and capture sales growth from the industry's technology transition.

2. Risk warning: Leading brands have already begun building underlying technical moats a decade in advance, and the pace of industry technology iteration will accelerate significantly. Sellers that fail to keep up with the transition and communicate clear intelligent value to consumers will likely fall far behind competitors, so they need to adjust their marketing focus promptly.

3. Actionable best practices: Li Auto has a proven track record of translating complex technology into specific use cases for family users, accurately hitting core consumer demand. Sellers can adapt this approach: convert abstract embodied intelligence technology into perceivable, scenario-based user experiences to improve conversion rates.

4. Shifting consumer demand: User demand for vehicles has already evolved from focusing on space and features to prioritizing intelligent experience. Sellers need to adjust their product selection and recommendation logic promptly to align with these new user expectations.

This article outlines the direction for intelligent transformation and new business opportunities for automotive manufacturing plants. Key insights are as follows:

1. Shifting product and production requirements: Vehicles in the embodied intelligence era impose far higher requirements on manufacturing, including higher-precision steer-by-wire chassis, actuators, and more reliable hardware integration. Safety and consistency requirements are also much stricter than for traditional models, so factories need to adjust their production standards quickly to adapt to new requirements.

2. Lessons for digital transformation: Li Auto has already used AI to restructure its R&D workflow, embedding AI across all business links and dramatically shortening product iteration cycles. Factories can follow this example by integrating AI into production, management and R&D end-to-end, solving low cross-departmental collaboration efficiency and boosting overall R&D and production productivity.

3. New business opportunities: As the embodied intelligence industry develops, demand for core components such as sensors, steer-by-wire systems and computing hardware will continue growing. Longer term, manufacturing demand for general humanoid robots will also gradually emerge. Factories can prepare by expanding relevant production capacity, accumulating intelligent manufacturing experience, and unlocking new business growth points.

This article summarizes pain points and industry trends during automakers' embodied intelligence transformation, pointing out clear strategic directions for service providers serving the automotive industry. Key takeaways are as follows:

1. Industry development outlook: The new energy vehicle industry has transitioned from the traditional supply chain integration stage to a new phase focused on full-stack in-house development of core technologies. A growing number of automakers are transforming into embodied intelligence companies, driving sustained rapid growth in demand for related technical services and solutions, creating a large addressable market.

2. Core client pain points: During the transformation process, automakers face core challenges including difficulty breaking down traditional R&D system silos, integrating AI into business workflows, low cross-departmental collaboration efficiency, and a shortage of AI-skilled talent. Traditional single-point service models can no longer meet automakers' end-to-end transformation needs.

3. Solution development direction: Service providers can develop integrated solutions focused on AI R&D tools, organizational transformation consulting, AI talent development, and model training infrastructure. These solutions help automakers break down software-hardware R&D barriers faster, shorten R&D iteration cycles, and support their full-stack in-house development transformation goals.

4. Technology positioning direction: New service demand will continue to emerge in areas such as perception models, co-design of chips and models, full vehicle operating systems, and steer-by-wire system integration. Early布局 in these technologies will allow service providers to capture first-mover advantage.

This article summarizes new demand created by automaker transformation, operational adjustment directions and risks for automotive-related platform operators. Key insights are as follows:

1. New industry demand: As automakers pursue embodied intelligence transformation and launch a new generation of intelligent vehicles, they require platforms to adapt with new technology demonstration, content distribution and user outreach approaches. User attention to autonomous driving and embodied intelligence-related content continues to rise, so platforms need to adjust their operational logic promptly to meet this new demand.

2. Business development and growth direction: Leading automakers have already begun布局 next-generation technologies, and new embodied intelligence-enabled models will reach the market in waves. Platforms can partner early with transforming leading automakers to secure rights such as new model premieres and priority promotion, attracting high-end users interested in intelligent technology and boosting overall platform traffic and transaction volume.

3. Operational and management adjustments: User focus for smart vehicles has shifted from price and features to underlying intelligent experience. Platforms can adjust their content sections and vehicle evaluation systems to add more technical explanations and intelligent experience testing content, better meeting user information demand.

4. Risk mitigation: Embodied intelligence is still in an early development stage, and many concepts have not yet been scaled commercially. Platforms should avoid overhyping unproven concepts, guide users to focus on real product experience, and prevent user trust crises from concept-driven market bubbles.

This article presents the latest industry developments in the convergence of automotive and AI, offering high reference value for industry research. Key insights are as follows:

1. New industry development: As the smart EV industry grapples with homogenization and uncertainty, Li Auto has proposed embodied intelligence as a new track, clearly positioning automobiles as the first entry point for AI into the physical world. It outlines a complete industry roadmap from autonomous driving to general humanoid robots, connects the underlying capability logic of the two tracks, and opens up a new growth narrative for the automotive industry.

2. Exploration of new business and organizational models: Li Auto has broken away from the traditional automaker model of relying on supplier integration, and pioneered a new path of full-stack in-house development of core technologies, building long-term competitive moats through upfront high R&D investment. It has also explored a new model of using AI to restructure organizational processes, redefining talent standards and R&D systems for the AI era, providing a new case study for technology manufacturing transformation.

3. New research directions: A number of new questions merit further in-depth research: the commercial return logic of the high-investment full-stack in-house development model, the path to overcoming barriers such as cost, reliability and safety during embodied intelligence commercialization, and how AI can reshape the organizational systems of traditional manufacturing enterprises. These all open up new directions for industrial research.

Disclaimer: The "Quick Summary" content is entirely generated by AI. Please exercise discretion when interpreting the information. For issues or corrections, please email run@ebrun.com .

I am a Brand Seller Factory Service Provider Marketplace Seller Researcher Read it again.

自动驾驶只是上半场,理想汽车押注具身智能下半场。

过去几年,理想汽车最擅长的事情,是把复杂的技术包装进一个足够具体的家庭场景里。

增程是为了消除家庭用户的里程焦虑,冰箱、彩电、大沙发是为了重新定义一辆家庭SUV的使用边界,理想过去的成功很大程度上来自这种能力。

但全新理想L9 Livis的上市,透露出一个新的信号。理想这一次想讲的,已经不只是“家庭旗舰SUV如何再升级”。在李想最近关于具身智能的表达中,全新理想L9 Livis更像是一个阶段性样本,理想试图把过去围绕家庭场景建立的产品能力,进一步延伸到更底层的技术和组织体系中。

如果说上一阶段的理想,回答的是“怎样造一辆更懂家庭用户的车”,那么这一次,它想回答的问题变成了:当汽车具备更强的感知、判断和行动能力之后,一家车企的边界会被推到哪里?

李想把这个方向称为具身智能。他提出,“自动驾驶是具身智能的上半场,通用人形机器人是具身智能的下半场”。这句话听起来像一个宏大的技术判断,但放到理想自身的发展脉络中看,它其实指向一个更现实的问题。理想已经不满足于只做一款爆品车型,而是希望围绕芯片、模型、操作系统、感知系统和执行机构,建立一套面向未来十年的底层能力。

对具身智能行业来说,人形机器人仍处在较早期阶段,距离大规模进入家庭和开放环境,还需要跨越成本、可靠性、泛化能力和安全性的多重门槛。而汽车已经是一个成熟的万亿级产业,因此自动驾驶的突破,很可能成为具身智能产业走向规模化的前置试验场。

全新一代理想L9,正是这个转向的第一个集中落点。它也是理想第一次把自研芯片、感知模型、VLA司机大模型、线控底盘、主动悬架和操作系统,放进同一套产品逻辑里。

它仍然是一辆家庭SUV,也仍然要面对高端新能源市场最直接的销量竞争。但在产品背后,理想正在完成一次更深的身份转换,从一家擅长定义家庭用车需求的汽车公司,走向一家试图掌握“物理世界智能化”底层能力的科技企业。

理想为什么要讲“具身智能”

理解理想的具身智能战略,或许可以从产业逻辑开始切入。

过去十年,AI主要改变的是信息世界。文字、图像、代码、搜索、知识管理,都已经被大模型重新塑造。但物理世界的变化相对缓慢。人们日常生活中的环境交互,仍然依赖人来完成。

汽车正是AI进入物理世界的关键入口之一。相比人形机器人,汽车拥有更成熟的产业链、更明确的使用场景,以及更大规模的数据闭环。它天然具备“身体”:有传感器感知环境,有计算平台处理信息,有控制系统执行动作,有操作系统协调各个模块。从这个意义上说,一辆具备高阶自动驾驶能力的汽车,本身就是一种具身智能产品。

李想已经把具身智能的发展拆成了一条较清晰的产业路线图:上半场是自动驾驶汽车,下半场是通用人形机器人。

在自动驾驶汽车这个上半场里,理想将其划分为三个阶段:2018年至2023年,是L2辅助驾驶阶段;2023年至2028年,是L3自动驾驶阶段;2028年至2033年,是L4无人驾驶阶段。

到了下半场,通用人形机器人又会进入新的三个阶段:2030年至2035年,具备相当于6岁儿童的泛化能力;2035年至2040年,达到12岁水平;2040年到AGI实现前后,具备接近18岁成人的泛化能力。

按照理想自己的技术划分,2023年至2028年的L3阶段,对应的是2D ViT感知、预训练模型、端到端控制,以及约2000TOPS级别算力;而2028年至2033年的L4阶段,才进一步走向3D ViT感知、稳定的预训练模型、全线控系统,以及接近10000TOPS级别算力。

这样来看,上半场积累的感知、模型、芯片、操作系统、控制能力,未来都可能成为下半场机器人的底层能力。理想将自动驾驶汽车和通用人形机器人都视为具身智能产品的核心形态,并判断未来L4自动驾驶用户与通用人形机器人用户存在较高重合度。

所以,理想这一次真正值得关注的地方,并不只是“做了自研芯片”、“做了线控底盘”这么简单,而是这些技术能力已经部分超过了当前阶段的基础要求,它代表着理想对未来增长曲线的重新设计。

战略背后的组织能力调整

如果只看产品和技术,理想的战略动作容易被理解为“加码AI研发”。但从前不久李想与罗永浩的对谈内容看,理想更深层的变化,是用AI重构组织和生产流程。

李想在对谈中多次提到,过去两百多天最重要的事情是学习AI。他不仅自己使用AI工具,也在公司内部推动员工使用Claude Code、OpenClaw等工具,并通过培训和分享让AI进入真实工作流程。

李想并不完全认同“一个人公司”的概念。他认为,建立稳定生产环境很难,AI并不会凭空创造东西,而是附着在真实业务流程中,提高研发、协同和交付效率。这也是理想内部推动AI的关键逻辑,他不把AI当成个人效率工具,而是把它嵌入研发、运营、产品和管理链条。

这对一家车企尤其重要。汽车是复杂工业品,涉及硬件、软件、供应链、制造、质量、安全、渠道和服务。单点效率提升固然重要,但真正困难的是跨部门协同。

理想在2026年年初完成研发体系变革,从按软硬件功能划分,转向按照“造具身智能”的方式重构,并打通传统研发中软硬件团队之间的壁垒;变革后,智能辅助驾驶模型训练迭代周期从两周缩短至一天。

可见,智能电动车竞争正在从产品定义能力,转向组织迭代能力。

李想对AI人才的判断也值得注意。他在对谈中提到,公司内部token消耗量靠前的人,并不一定是过去标准下最顶级的员工;有些人表达能力不强、过去获取资源有限,但“脑子极强”,一旦有token和业务环境,就可以改造很多东西。

这说明理想正在重新定义组织里的生产力。过去企业看重汇报能力、管理层级和资源协调;AI时代,真正能用AI改造业务流程、构建生产环境、完成闭环的人,可能成为新的关键人才。

所以,理想从车企向具身智能企业的转型,不能只理解为产品转型。它更是一场组织工程:把人、AI工具、业务流程和技术平台重新组合起来。最终目标不是用AI替代人,而是让组织拥有更高密度的创造力和执行力。

全栈自研是具身智能的入场券

理想的具身智能战略,最容易被质疑的一点是:一家车企为什么要做这么重的全栈自研?

在传统汽车工业中,车企的核心能力很大程度上是定义产品、整合供应链、制造和渠道能力。发动机、变速箱、底盘、电控等,各个模块都有成熟供应商体系。车企可以通过采购和集成,快速推出有竞争力的产品。

但具身智能的难点在于,它不是单个模块的智能,而是一个系统如何实时理解世界并作用于世界。感知、模型、算力、操作系统和执行机构之间的协同效率,会直接决定车辆在复杂场景中的表现。

李想用人体来类比这套系统:芯片是心脏,模型是大脑,感知系统是眼睛,底盘是手脚,操作系统是神经系统。具身智能产品的能力上限,取决于整套系统能否被统一设计、统一调度、统一迭代。

这正是理想坚持全栈自研的底层原因。

以芯片为例,理想自研的5纳米马赫M100芯片,采用数据流架构,双颗算力达到2560TOPS。但对理想来说,自研芯片的意义并不只是“算力更高”,而是芯片和模型可以协同设计。未来智能驾驶模型越来越复杂,如果芯片架构与模型需求不匹配,就可能出现算力很高,但实际效率不足的问题。

再看操作系统。星环OS的价值,是想成为整车智能体的“神经系统”。当车辆进入全线控阶段,转向、制动、悬架等执行机构都需要被AI实时调度,系统延迟、安全冗余和控制精度就会成为核心问题。这意味着,具身智能时代的车企竞争,核心在于把感知、决策和控制做成一个可靠闭环。

从产品角度看,全新一代理想L9 Livis最重要的变化,并非外观、座舱或者舒适配置,而是它第一次把理想所说的具身智能五大核心能力集中落地在一辆量产车上。

感知层面,L9 Livis从2D ViT进化到3D ViT,并将激光雷达的三维几何信息与摄像头语义信息在编码阶段统一。模型层面,理想引入马赫VLA模型。控制层面,搭载“完全体”线控底盘和800V主动悬架。这组技术放在一起看,理想想要证明“AI控制物理世界”的能力已经开始具备硬件基础。

这也是理想全栈自研的商业逻辑。短期看,全栈自研投入重、周期长、风险高;但长期看,它有机会形成技术代差和产品差异化。数据显示,理想汽车2025年研发投入达113亿元,其中AI相关投入占50%;2026年研发投入预计仍保持在120亿元左右,AI相关研发投入占比约一半。

对理想来说,这相当于把未来十年的竞争壁垒前置建设。尤其在智能电动车行业进入同质化之后,只有掌握底层技术架构的企业,才有可能在体验、安全和成本之间建立更稳定的平衡。

当然,全栈自研并不天然等于成功。它需要持续高投入,也需要产品销量支撑现金流,更需要组织能力承接复杂研发。但李想在与罗永浩的对谈中说,“理想汽车做AI,不是冒险。不做才是冒险。” 这句话某种程度上概括了理想此轮战略转向的决心与笃定。

汽车正在成为物理世界AI的起点

理想选择在L9上集中呈现具身智能,并不意味着这套战略只服务于一款旗舰SUV。更长远看,它释放出的信号是:当智能电动车竞争进入同质化之后,汽车产业正在寻找下一轮增长叙事,而具身智能可能成为科技产业与汽车产业重新交汇的方向。

过去十年,中国汽车行业完成了从机械产品到智能终端的转变。但到今天,续航、空间、配置、补能和智能驾驶都在快速趋同。行业真正的迷茫在于,下一步到底靠什么拉开差距。

理想给出的答案,是把汽车放进更大的科技产业坐标中重新理解。

在李想的判断里,自动驾驶解决的是机器如何在道路环境中感知、判断和行动;人形机器人要解决的,则是机器如何进入家庭、工厂和更复杂的开放场景。二者看似形态不同,本质上都指向同一个问题:AI如何理解真实世界,并对物理世界产生稳定、可靠、可控的影响。

这也是理想重投芯片、模型、操作系统和线控底盘的原因。它们短期服务于汽车,长期则可能成为进入更多智能终端的底层能力。汽车是第一个成熟载体,因为它有高价值硬件、明确场景、持续数据反馈和足够强的用户支付能力。但汽车未必是终点,它更像是物理世界AI最先规模化的训练场。

当然,具身智能不是一个靠概念就能兑现的赛道。全栈自研意味着高投入,机器人业务也需要更长周期验证。市场最终看的,仍然是产品体验、安全可靠性和商业效率。汽车可以成为物理世界AI的入口,但前提是它必须先是一辆足够好、足够稳定、足够可信的车。

全新一代理想L9释放出的信号已经足够清晰:在行业普遍陷入“下一步卷什么”的迷茫时,理想试图把汽车竞争从单车产品力,推向更底层的系统能力竞争,今后真正的竞争可能会变成,谁能更早建立起连接感知、判断与执行的完整系统。

自动驾驶只是开始。汽车产业真正的新变量,是它正在成为AI进入物理世界的第一站。而理想想争夺的,也不只是一款旗舰SUV的市场份额,而是下一轮科技产业与汽车产业融合的入场券。

注:文/李玉鹏,文章来源:钛媒体(公众号ID:taimeiti),本文为作者独立观点,不代表亿邦动力立场。

文章来源:钛媒体

广告
微信
朋友圈

这么好看,分享一下?

朋友圈 分享

APP内打开

+1
+1
微信好友 朋友圈 新浪微博 QQ空间
关闭
收藏成功
发送
/140 0