6月29日,国内自动驾驶解决方案提供商Momenta赴港招股,公布了14家基石投资者,比亚迪、梅赛德斯-奔驰两家车企共投资4000万美元(约合人民币2.7亿元),其中比亚迪投资1500万美元,梅赛德斯-奔驰投资2500万美元。
(图源:Momenta)
Momenta是与华为、大疆(卓驭)、地平线并列「华大地魔」的国内四大自动驾驶企业之一,更是被誉为「物理AI第一股」。
可想而知,接下来车企和供应商之间不会再有那么多暧昧的试探,站队会越来越明确,行业资源也会快速向这几家头部垄断,留给其他玩家的时间真的不多了。
整个行业也将在这一过程中,逐渐实现格局定型。
智驾卷入大模型时代,“回血”成了头等大事
事实上,Momenta并非「华大地魔」四家企业中第一个上市的,早在2024年10月24日地平线就在港交所上市。今年一季度,卓驭CEO沈劭勋也在接受路透社专访时透露,公司最快 2027 年赴港上市。
而华为拆分汽车业务成立引望智能,在今年年初也曾有消息称其计划2026年下半年上市,虽说引望智能已于1月11日官方辟谣。不过在电车通(ID:dianchetong233)看来,华为将汽车业务剥离出来成立引望智能,又让赛力斯和阿维塔入股,上市估计只是时间为问题。
Momenta上市反映出的,是自动驾驶竞争进入白热化,资源投入门槛持续抬升。自动驾驶技术已迈入大模型时代,而大模型训练需要持续投入巨额资金。以Momenta为例,2025年研发投入高达18.69亿元,近三年累计研发投入达46.6亿元。在内存、闪存价格持续走高的背景下,自动驾驶企业的研发投入也逐年上涨。
(图源:Momenta)
正因如此,Momenta、地平线等仍在亏损的企业需要同时拉动投资和巩固与车企的关系,上市融资是它们最好的选择。通过发行股票不仅能够补充现金流,还可引入车企产业资本,加深双方合作绑定。
Momenta公布的招股书信息显示,一家2021年与其建立合作关系的中国电动车、动力电池制造公司是Momenta的最大客户,为其提供了21.6%的营收。结合Momenta披露的14家基石投资者名单,其中仅比亚迪、梅赛德斯-奔驰两家车企,由此可判断其最大客户为比亚迪。
(图源:Momenta)
Momenta通过获得比亚迪的投资与后者建立更深层次的合作关系后,有望进入更多比亚迪车型。
地平线与Momenta处境相似,因而早在2024年就在港股上市。卓驭背靠大疆,资金储备充足,上市诉求不强,预计明年或后年上市;华为自身资金实力雄厚,鸿蒙智行已与赛力斯、江汽、北汽、上汽、奇瑞等多家车企深度合作,旗下引望智能几乎不存在上市刚需,具体IPO规划有待进一步披露。
Momenta的上市,代表着行业进入终局分化阶段,行业将形成头部企业主导、多层级主体共存、新入局者难以入场的稳定格局。技术、资金实力薄弱的自动驾驶企业将加速出清,产业资源持续向头部龙头与专精特新企业集中。
华为走全栈自研、构筑高端技术壁垒路线;卓驭主打极致轻量化、低成本,深耕垂直落地场景;地平线聚焦算力底座与核心硬件,方案通用性强;Momenta则以世界模型、物理 AI 为核心,打造开放合作平台,四家企业均在各自赛道持续深耕。
当下称其为行业垄断巨头尚且为时过早,但「华大地魔」四家企业已显现龙头发展潜力,并且都有专精领域,有机会在自动驾驶领域新阶段成长为行业领导者。至于国内其他自动驾驶企业,也需要尽快融资,才能避免在更激烈的竞争中掉队。
受Momenta影响的不只有自动驾驶企业,面对Momenta的上市,车企实际上也站在了「人生的十字路口」。
一年烧掉几十亿,车企开始妥协
观察者网《2024智驾年度调查》显示,83.28%消费者称高阶智驾会左右购车选择。时隔一年多,国内消费者对高阶智驾的需求只增不减。
为迎合消费者的要求,比亚迪、零跑等车企将智驾平权当作重中之重,不断下放高阶智驾。如今10万元以内价位,消费者就能买到搭载激光雷达的比亚迪海鸥、零跑A10 等车型。
然而在高阶智驾加速普及之际,车企却普遍面对着一个问题——该不该自研智驾技术。
关于车企智驾技术应该自研还是外部采购的话题,行业已经讨论过无数次了。当前主流车企均布局智驾自研,但短期内自研技术难以比肩头部车企与专业自动驾驶公司。因此多数车企采取分层路线,低端车型搭载自研智驾系统,高端车型暂采用第三方外采方案。
(图源:Momenta)
但正如电车通(ID:dianchetong233)前文所言,自动驾驶自研投入成本高昂,企业还要应对车规认证、安全合规两大难题,对车企而言是沉重负担。如比亚迪一般,与自动驾驶企业Momenta深度绑定,成立合资公司迪派智行,在力所能及的范围内自研域控制器、芯片等硬件,软件方面则选择Momenta的算法,或许是车企最好的选择。
高阶智驾落地速度,将直接决定车企市场份额走势。Momenta 等第三方智驾方案商成熟的技术,让中小车企也能快速搭载高阶智驾,行业竞争重心从产能规模比拼,转向车载智能体验竞争。
电车通(ID:dianchetong233)认为,Momenta上市后,部分智驾技术储备薄弱的车企,会适度缩减自研投入,通过战略入股、合资成立子公司等方式,与第三方智驾方案商开展合作,借力第三方企业的技术方案,快速提升产品的智驾体验。
Momenta之所以被称为「物理AI第一股」,原因在于它是全球首个以「物理AI基座模型」(世界模型)为核心定位,并成功登陆公开资本市场的独立公司,且在自动驾驶这一物理AI核心场景中,率先实现了技术落地、数据闭环、商业规模化的完整验证。
(图源:Momenta)
物理AI的核心在于数据规模化与商业落地规模化。若车企全部选择全栈自研,各企业数据相互割裂,商业化推进受阻,不仅会制约物理AI技术迭代,也将延缓国内汽车产业从传统制造向「AI 定义汽车」转型的进程。
在此背景下,车企弱化全面自研、拥抱产业协同已成必然趋势。这种轻量化、产业协同模式,既能帮助车企规避高额自研风险、缩短智驾技术迭代周期,也能打破行业数据孤岛,完善物理AI产业生态。Momenta的资本化落地,恰好为行业提供了成熟的合作范本,持续推动国内汽车产业摆脱传统造车逻辑,全面迈入AI驱动的智能化全新发展阶段。
不卖“灵魂”卖协同,物理AI靠什么讲新故事?
Momenta招股,其实是给国内智驾行业的混战拉下了大幕。第一梯队的“四大天王”已经把核心资源和赛道分得差不多了,留给中小智驾企业的窗口期正在死死关闭。
现在的行业格局很有意思,四大头部玩家各自拿出了看家本领:华为技术强硬、大包大揽;地平线深耕硬件、做底层基座;大疆卓驭深谙降本,死磕垂直场景;而Momenta则靠着世界模型和数据闭环,成了车企最放心的“技术合伙人”。
过去很多车企对第三方供应商充满戒备,但奔驰和比亚迪这次重仓Momenta,给行业打了个样:车企出钱出车提供数据,智驾公司出技术加速迭代,大家各司其职。
这种互不越界的合作,远比车企自己硬撑着烧钱、或者供应商两眼一抹黑死磕,要现实得多。
(图源:Momenta)
随着Momenta递表,全栈自研的红利期已成为过去,高昂的研发成本正在把中小玩家生生拖垮。现在的行业共识变得很现实:能外采的绝不瞎折腾,用股权换技术,快速把智驾体验补齐才是正道。
这也解释了为什么Momenta能成为国际车企巨头的“香饽饽”。合资车企在国内要面对复杂的数据合规和技术封锁,自己做根本来不及,直接买本土成熟的方案是最理性的选择。奔驰、宝马、奥迪纷纷给Momenta买单,就是最直接的证明。
在这场由物理AI驱动的智能化洗牌中,车企弱化自研、转向量产协同已经不是“愿不愿意”的问题,而是在智能化下半场中的唯一选题,也是最好的选择。
End解锁车主专属圈,同系车主,都在这里!入群畅聊爱车,欢迎大家扫码添加小助理拉群!
Author: Shīhún Yǐn | Supervisor: Luó Chāo
On June 29, the domestic autonomous driving solution provider Momenta went public in Hong Kong and announced 14 cornerstone investors, including BYD and Mercedes-Benz, which invested a total of $40 million (approximately RMB 270 million). Specifically, BYD invested $15 million and Mercedes-Benz invested $25 million.
(Image source: Momenta)
Momenta is one of China's four major autonomous driving enterprises, alongside Huawei, DJI (Zhuoyu), and Horizon Robotics. It is also known as the "first physical AI stock."
It is evident that the ambiguous probing between automakers and suppliers will diminish in the future, with alignments becoming increasingly clear. Industry resources will rapidly concentrate toward these leading companies, leaving little time for other players. The industry landscape will gradually stabilize through this process.
In fact, Momenta is not the first of the "four major enterprises" to go public. Horizon Robotics was already listed on the Hong Kong Stock Exchange on October 24, 2024. Earlier this year, Zhuoyu's CEO, Shen Shaoxun, stated in an interview with Reuters that the company plans to go public in Hong Kong as early as 2027.
Regarding Huawei, there were reports in early 2024 suggesting that the spin-off of its automotive business, Ruwan Intelligence, was planning for an IPO in the second half of 2026. Although Ruwan Intelligence officially denied this on January 11, it is widely believed that listing is only a matter of time given its significant investments from companies like Seres and Avatr.
Momenta's IPO reflects the intensification of autonomous driving competition and the rising barriers to entry. Autonomous driving technology has now entered the era of large models, which require substantial financial investment for training. For instance, Momenta's R&D expenditure reached RMB 1.869 billion in 2025, with a three-year cumulative total of over RMB 4.6 billion. Given the continuous rise in memory and flash storage prices, R&D costs for autonomous driving companies are expected to remain high.
Because of these high costs, companies like Momenta and Horizon Robotics need to secure funding while also strengthening ties with automakers. Going public is the ideal solution as it not only provides cash flow but also allows for the introduction of industrial capital from partner automakers, deepening collaborative bonds.
The prospectus indicates that one of Momenta's largest clients, an electric vehicle and battery manufacturer partnered since 2021, contributed 21.6% of its revenue. Analyzing the list of cornerstone investors—consisting mostly of car manufacturers—it is highly likely that BYD is the primary client referenced, as both BYD and Daimler were mentioned prominently.
By securing BYD's investment and establishing a deeper partnership, Momenta is positioned to expand its presence across a wider range of BYD models.
This pattern is consistent with Horizon Robotics' strategy. While Huawei, backed by strong cash reserves and a well-established distribution network, may not prioritize immediate listing, Momenta's successful IPO signifies that the industry is entering a phase of consolidation. We can expect to see further mergers and acquisitions or strategic partnerships between autonomous driving providers and automakers.
The listing represents the beginning of the end for the era of "cheap and fast" full-stack self-development. With the high computational demands of AI training, even major automakers are finding it cost-effective to outsource to specialized providers rather than build everything in-house.
Automakers are beginning to compromise.
A survey by Observer Net shows that 83.28% of consumers consider high-level autonomous driving a key factor in their purchasing decisions. As demand for advanced driver-assistance systems (ADAS) increases, particularly in lower-priced segments where manufacturers like BYD and Leapmotor are extending laser radar availability, the question of whether to self-develop or purchase technology has become critical.
While many automakers have invested in their own ADAS capabilities, the technology is often outpaced by specialized companies and lacks the maturity of leading vendors. Therefore, most are adopting a tiered approach: using their own systems for basic features and purchasing higher-level algorithms from partners like Momenta for premium models.
(Image source: Momenta)
The time-to-market for high-level autonomous driving is crucial. By partnering with mature providers like Momenta, manufacturers can quickly integrate advanced features into their product lines. This allows them to compete on the quality of the driving experience and advanced functions rather than just on production capacity.
In response to Momenta's success, some automakers with less-developed technology are likely to reduce independent R&D investment and instead pursue strategic partnerships or joint ventures, leveraging external expertise to rapidly improve their vehicle offerings.
(Image source: Momenta)
The transition from a fragmented industry to a collaborative ecosystem is clearly unfolding. As data becomes the limiting factor, automakers must work with companies like Momenta to pool datasets and accelerate the advancement of physical AI. This shift from a "gold rush" mentality to a "data-sharing" model is the necessary evolution for the automotive industry.
But as Dianchetong (ID: dianchetong233) previously mentioned, the cost of self-developed autonomous driving is high, and companies must also contend with automotive certification and safety compliance, which is a heavy burden for automakers. Like BYD, partnering deeply with an autonomous driving company like Momenta and establishing a joint venture (Dipai Zhisheng) is likely the best choice for automakers, allowing them to focus on hardware like domain controllers and chips while leveraging Momenta’s algorithms for software.
The speed at which advanced intelligent driving is rolled out will directly determine an automaker’s market share. Mature technologies from third-party providers like Momenta enable even small and medium-sized automakers to quickly adopt advanced autonomous driving, shifting the focus of industry competition from scale to onboard intelligent experience.
Dianchetong (ID: dianchetong233) believes that after Momenta’s IPO, some automakers with weak autonomous driving technology reserves will reduce their self-development efforts and instead collaborate with third-party providers through strategic investments or joint ventures to quickly improve their product’s intelligent driving experience.
Momenta is known as the “first physical AI stock” because it is the first independent company globally to be listed on a public capital market with “a physical AI foundation model” (world model) as its core, and has successfully demonstrated the commercialization of this technology in the critical area of autonomous driving.
(图源:Momenta)
The core of physical AI lies in large-scale data collection and commercialization. If all automakers were to self-develop, data would be fragmented, and commercialization would be stifled, hindering the advancement of physical AI technology and slowing the transition of the automotive industry to being defined by AI.
In this context, automakers reducing self-development and embracing industry collaboration is a necessity. This lightweight, collaborative model helps automakers avoid high R&D risks, shortens the iteration cycle, breaks down data silos, and completes the physical AI industry ecosystem. Momenta’s successful public offering provides a mature model for this cooperation, driving the domestic automotive industry toward a new stage of AI-driven development.
Selling Synergy, Not Souls: How Does Physical AI Tell a New Story?
Momenta’s IPO effectively draws a line under the chaotic battle for autonomous driving in China. The top four players have already carved up the core resources and markets, leaving a very narrow window for new entrants.
The current landscape is well-defined: Huawei excels in technology and integration; Horizon specializes in hardware and foundational platforms; DJI’s Zhuoyu focuses on cost-effective, vertical-specific solutions; and Momenta has established itself as the most trusted “technology partner” for automakers through its world model and data-centric approach.
While automakers were once wary of third-party suppliers, the investments from Mercedes-Benz and BYD validate that the best strategy is for automakers to provide data and hardware while specialized companies handle the complex, high-cost algorithm development.
This collaborative model is far more efficient than automakers attempting to “boil the ocean” by developing everything themselves, or suppliers operating in isolation.
(图源:Momenta)
As Momenta goes public, the era of high-cost, full-stack self-development is ending. The industry now recognizes that companies should focus their resources where they are most competitive and outsource the rest.
Consequently, Momenta is attracting significant interest from international automakers who face restrictions on local data collection and require ready-made solutions to compete in the Chinese market.
In a landscape reshaped by physical AI, automakers are shifting from full self-reliance to strategic partnership. This is not just the easy choice; it is the necessary one to survive in the intelligent driving era.
#Momenta #Huawei #Zhuoyu #AutonomousDriving #Horizon
End Unlock the exclusive owner circle, where fellow car owners gather! Join the group to chat about your vehicle — scan the code to add the little assistant and get pulled into the group!








