广告
加载中

上海、北京国资 联手押注一家AI4M公司

韦香惠 2026-07-14 09:50
韦香惠 2026/07/14 09:50

邦小白快读

EN
全文速览

本文核心信息是当前AI for Materials(AI4M,材料人工智能)作为科学人工智能(AI4S)的核心细分赛道进入爆发期,国内头部AI4M创业公司深度原理获得近10亿元A系列融资,获得上海、北京两地国资联手押注,核心干货如下:

1. 当前全球AI4S已经成为资本追逐的新高地,今年以来全球该领域融资规模超44亿美元,AI4M作为AI连接实体产业的重要入口,是全球资本竞相布局的新赛道,海外多家AI4M企业已经获得贝佐斯家族办公室、英伟达等头部资本投资,估值短期内大幅上涨。

2. 深度原理是国内目前拿到AI4M领域最大规模融资的企业,核心竞争力在于打通了材料研发全流程,解决了AI计算和实验验证脱节的行业痛点,目前已经实现商业化落地,服务多家头部企业,落地多个单客户千万级项目。

对于布局新材料相关业务的品牌商,本文透露出AI4M赛道的发展趋势、产品研发机会以及产业风向,干货整理如下:

1. 政策与产业风向层面,AI驱动材料研发已经被纳入国家战略布局,国务院将人工智能赋能科技创新列为培育新质生产力的重要方向,是我国建设世界科技强国的重要战略机遇,下游产业资本已经认可AI材料研发的商业价值,赛道进入加速发展期。

2. 产品研发层面,AI4M技术可以大幅压缩材料研发周期,将原本需要数天甚至数月的计算压缩到0.4秒,还能实现全研发流程自动化,帮助品牌商降低研发成本、缩短新品上市周期。品牌商可选择和AI4M企业合作,通过管线共研、里程碑付款的模式开展合作,目前已有多个千万级合作项目落地。

3. 消费端相关的营养日化、精细化工等领域都已经有AI4M技术落地,未来会推动更多新材料产品问世,品牌商可提前布局相关技术应用。

对于关注AI、新材料赛道的卖家,本文梳理了AI4M赛道的政策动向、市场机会以及风险提示,干货整理如下:

1. 政策层面,国家明确将人工智能赋能科技创新列为培育新质生产力的核心方向,AI4M作为AI落地实体产业的核心入口,属于政策重点支持的新兴增长赛道,获得京沪两地国资联手布局,赛道确定性持续提升。

2. 市场机会层面,AI4M目前仍处于早期发展阶段,尚未形成广泛市场共识,但具备巨大的想象空间,下游新能源、电子材料、精细化工、营养日化等多个领域都有旺盛需求,商业化路径已经跑通,可通过订阅服务、管线共研、成果分成等方式实现盈利,已有千万级项目落地。

3. 风险提示方面,AI4M商业化周期长,对产业理解要求高,单纯追求技术先进但未对准真实产业需求的项目大概率会失败,进入该领域需要先找准产业痛点再落地技术。

对于从事材料生产、新品研发的工厂,本文带来了AI赋能研发的新方向以及商业机会,干货整理如下:

1. 产品生产和研发需求层面,传统材料研发存在周期长、试错成本高、干实验与湿实验脱节的痛点,AI4M技术可以将研发环节的计算时间从数月压缩到0.4秒,还能打通AI预测、实验验证、模型优化的全流程闭环,实现研发自动化,能帮助工厂大幅提升研发效率,降低研发成本。

2. 商业机会层面,AI4M赛道获得国资、产业资本、市场化投资机构的共同押注,是国家战略支持的新兴方向,在新能源、半导体显示、化工、合金材料等多个领域都有巨大的应用需求,工厂可以和AI4M企业开展研发管线共研,合作开发新材料,共享研发成果,拓展新品类。

3. 数字化转型启示,工厂可对接成熟的AI研发平台,借助外部技术能力推进自身研发数字化,不用从零搭建全套AI技术体系,降低转型门槛。

对于服务AI、新材料产业的服务商,本文梳理了AI4M赛道的发展趋势、行业痛点以及可参考的解决方案,干货整理如下:

1. 行业发展趋势方面,AI for Science已经被英伟达列为人工智能三大发展方向之一,AI4M作为AI4S的核心细分赛道,今年全球AI4S融资规模已经超过44亿美元,该赛道已经从资本市场的前沿探索逐步进入国家战略布局视野,未来会释放大量的技术服务、研发服务需求,行业增长空间巨大。

2. 行业核心客户痛点:当前AI4M行业普遍存在三个痛点,一是多数企业仅覆盖单点研发环节,没有打通从设计到落地的全流程;二是AI计算(干实验)和实验室验证(湿实验)脱节,依赖人工沟通,模型无法持续迭代优化;三是很多项目技术先进但没有对准产业真实需求,难以商业化落地。

3. 可参考的解决方案:可参考头部创业公司的路径,打造覆盖全研发链条的一体化解决方案,建立AI预测到模型优化的完整闭环,从起步阶段就围绕产业真实需求设计研发方向,提升落地成功率。

对于聚焦硬科技、AI赛道的投资平台、产业平台,本文梳理了AI4M赛道的平台需求、布局动向以及风险规避方向,干货整理如下:

1. 行业对平台的需求:AI4M属于长周期投资赛道,单一类型资本难以满足其发展需求,当前行业已经验证了“国资+产业资本+市场化VC”的组合投资模式,国资提供长期稳定的资金支持,产业资本对接产业需求,市场化机构提供运营资源,三方互补能更好支撑项目发展。

2. 平台最新布局动向:上海、北京两地的国资平台已经率先联手布局AI4M赛道,释放出AI驱动科学创新进入国家战略布局的明确信号,地方国资平台可围绕AI硬科技赛道设立专业化产业基金,提前布局优质项目。

3. 风险规避方向:AI4M商业化周期长,短期难以实现规模化收入,平台布局该赛道需要做好长期投资的准备,优先选择有产业背景、对准真实需求、已经实现商业化落地的项目,规避只有技术展示没有落地场景的项目。

对于研究AI产业、硬科技发展的研究者,本文梳理了AI4M赛道的最新产业动向、新问题以及商业模式、政策层面的启示,干货整理如下:

1. 产业新动向:当前AI4S已经成为全球资本追逐的新高地,今年以来全球融资规模已经超过44亿美元,AI4M作为AI连接实体产业的核心入口,成为全球资本布局的新热点,国内AI4M赛道迎来爆发,国内头部项目深度原理获得近10亿元融资,是国内AI4M领域迄今最大规模融资,京沪国资联手投资,标志着AI驱动科学创新从资本市场前沿探索进入国家战略布局视野。

2. 行业新问题:AI4M目前仍是相对非共识的赛道,存在商业化周期长、研发验证成本高、干实验湿实验脱节、技术与产业需求错配等问题,多数创业公司停留在技术展示阶段,实现规模化收入的企业较少。

3. 商业模式与政策启示:商业化可采用单点问题解决、AI Agent订阅、研发管线共研结合里程碑付款、成果分成的多元模式;政策层面需要鼓励长周期硬科技投资,支持多类型资本联合布局,推动AI技术落地实体产业。

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 outlines that AI for Materials (AI4M), a core subfield of AI for Science (AI4S), is currently entering a period of explosive growth. Leading Chinese AI4M startup Deep Principle has recently closed a Series A financing round totaling nearly 1 billion yuan, backed by a joint investment from state-owned investors in Shanghai and Beijing. Key takeaways are as follows:

1. AI4S has emerged as a new hot spot for global capital, with total financing in the sector exceeding $4.4 billion globally so far this year. As a critical entry point for AI to connect with the real economy, AI4M has become a new key track for global investors. Multiple overseas AI4M companies have already secured backing from top investors including the Bezos Family Office and NVIDIA, with valuations surging sharply in a short period.

2. Deep Principle is currently the largest-funded AI4M startup in China. Its core competitive advantage lies in its end-to-end integration of the entire material R&D workflow, which solves the long-standing industry pain point of disconnected AI computing and experimental validation. The company has already achieved commercialization, serving multiple leading industry clients and delivering multiple 10 million-yuan level projects for single customers.

For brand owners with new material-related business布局, this article outlines AI4M's development trends, R&D opportunities and industry outlook. Key insights are as follows:

1. On policy and industry outlook: AI-driven material R&D has been incorporated into China's national strategic布局. The State Council has identified AI-enabled scientific innovation as a core driver for cultivating new quality productive forces, representing a key strategic opportunity for China to build itself into a global science and technology power. Downstream industrial capital has already recognized the commercial value of AI-powered material R&D, and the sector is entering a phase of accelerated growth.

2. On product R&D: AI4M technology dramatically cuts material R&D cycles, compressing calculation tasks that originally took days or even months down to 0.4 seconds, while enabling full end-to-end R&D automation. This helps brand owners reduce R&D costs and shorten time-to-market for new products. Brands can partner with AI4M companies via joint R&D pipelines with milestone-based payment structures, and multiple 10 million-yuan level cooperation projects have already been delivered.

3. AI4M has already been deployed in consumer-facing sectors including nutrition, personal care and fine chemicals, and will drive the launch of more new material products in the future. Brands can布局相关 technology applications in advance.

For sellers focused on the AI and new materials sectors, this article summarizes AI4M's policy trends, market opportunities and risk warnings. Key takeaways are as follows:

1. On policy: China has explicitly identified AI-enabled scientific innovation as a core direction for cultivating new quality productive forces. As a key entry point for AI to落地 in the real economy, AI4M is a high-priority emerging growth track supported by national policy, with joint布局 by state-owned investors from Beijing and Shanghai boosting the sector's growth certainty.

2. On market opportunities: AI4M remains in an early development stage and has not yet reached broad market consensus, but it offers enormous growth potential. It boasts strong demand across multiple downstream sectors including new energy, electronic materials, fine chemicals, and nutrition and personal care. Commercial pathways have already been proven viable, with revenue achievable via subscription services, joint R&D pipelines, and profit sharing from R&D outcomes, and multiple 10 million-yuan level projects have already been delivered.

3. On risk warnings: AI4M has a long commercialization cycle and requires deep industry understanding. Projects that pursue technical advancement alone without addressing real industry demand are very likely to fail. Entrants into the sector should identify industry pain points first before deploying technology.

For factories engaged in material production and new product R&D, this article outlines new directions and commercial opportunities brought by AI-enabled R&D. Key insights are as follows:

1. On production and R&D needs: Traditional material R&D suffers from long cycles, high trial-and-error costs, and disconnection between dry lab (computation) and wet lab (experimental work). AI4M compresses R&D computation time from months to 0.4 seconds, and enables a closed full loop covering AI prediction, experimental validation and model optimization to automate R&D. This helps factories drastically improve R&D efficiency and cut R&D costs.

2. On commercial opportunities: AI4M has secured joint backing from state-owned capital, industrial capital and market-oriented investment institutions, and is an emerging direction supported by national strategy. It has enormous application demand across sectors including new energy, semiconductor display, chemicals, and alloy materials. Factories can conduct joint R&D pipelines with AI4M companies to co-develop new materials, share R&D outcomes and expand new product categories.

3. Implications for digital transformation: Factories can connect with mature AI R&D platforms to advance their own R&D digitalization via external technical capabilities, rather than building a full set of AI systems from scratch, which lowers the barrier to transformation.

For service providers serving the AI and new materials industries, this article summarizes AI4M's development trends, core industry pain points and reference solutions. Key takeaways are as follows:

1. On industry development trends: NVIDIA has listed AI for Science as one of the three core development directions for artificial intelligence. As a core subfield of AI4S, AI4M is part of a global AI4S sector that has raised over $4.4 billion in financing this year. The sector has evolved from a frontier exploration area for capital markets into a key focus of national strategic布局, and will release massive demand for technical services and R&D services in the future, with enormous industry growth potential.

2. Core pain points for industry clients: The AI4M sector currently faces three widespread pain points: first, most companies only cover single points of the R&D process, and have not built out end-to-end workflows from design to commercialization; second, AI computing (dry experiments) is disconnected from laboratory validation (wet experiments), relying on manual communication and preventing continuous model iteration and optimization; third, many projects are technically advanced but misaligned with real industry demand, making commercialization difficult.

3. Reference solutions: Players can follow the path of leading startups, build integrated solutions covering the full R&D chain, establish a complete closed loop from AI prediction to model optimization, align R&D directions with real industry needs from the early stage, and improve the success rate of commercial落地.

For investment platforms and industrial platforms focused on hard tech and AI tracks, this article summarizes AI4M's demand for platforms, latest布局 trends and risk mitigation directions. Key takeaways are as follows:

1. Industry demand for platforms: AI4M is a long-cycle investment track, and a single type of capital cannot meet its development needs. The industry has already validated the "state-owned capital + industrial capital + market-oriented VC" combined investment model: state-owned capital provides long-term stable capital support, industrial capital connects to industry demand, and market-oriented institutions provide operational resources. This complementary tripartite structure better supports project development.

2. Latest platform布局 trends: State-owned platforms from Beijing and Shanghai have already taken the lead in jointly布局 the AI4M track, sending a clear signal that AI-driven scientific innovation has entered national strategic布局. Local state-owned platforms can set up specialized industrial funds focused on AI hard tech to布局 high-quality projects in advance.

3. Risk mitigation directions: AI4M has a long commercialization cycle and it is difficult to generate scalable revenue in the short term. Platforms布局 this track need to prepare for long-term investment, and prioritize projects with industry background, alignment with real demand and existing commercial落地, while avoiding projects that only have technical demos without real落地 scenarios.

For researchers studying AI industry and hard tech development, this article summarizes the latest industry trends, new emerging issues, and implications for business models and policy in the AI4M track. Key takeaways are as follows:

1. New industry trends: AI4S has become a new hot spot for global capital, with total global financing exceeding $4.4 billion so far this year. As a core entry point for AI to connect with the real economy, AI4M has become a new focus for global capital布局, and the domestic AI4M track is experiencing an outbreak. Leading domestic startup Deep Principle has secured nearly 1 billion yuan in financing, the largest ever raised in China's AI4M sector, in a joint investment by Beijing and Shanghai state-owned capital. This marks that AI-driven scientific innovation has moved from frontier exploration in capital markets into national strategic布局.

2. New industry issues: AI4M remains a relatively non-consensus track, facing challenges including long commercialization cycles, high R&D validation costs, disconnection between dry and wet experiments, and misalignment between technology and industry demand. Most startups remain stuck at the technical demonstration stage, with few companies achieving scalable revenue.

3. Implications for business models and policy: Commercialization can adopt a diversified model combining single-point problem solving, AI Agent subscriptions, joint R&D pipelines with milestone payments, and outcome-based profit sharing. On the policy side, authorities should encourage long-term investment in hard tech, support joint布局 by multiple types of capital, and promote the落地 of AI technology in the real economy.

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.

AI for Science(AI4S,科学人工智能)正在成为全球资本追逐的新高地。据不完全统计,今年以来,全球AI4S领域融资规模已超过44亿美元。

作为AI与物理世界的重要连接点,新材料研发被认为是人工智能突破数字空间、真正进入现实产业的重要入口,因此,AI for Materials(AI4M)也是全球投资机构在AI4S领域中竞相布局的新赛道之一。

放眼海外,资本早已率先押注这一趋势。英国公司Cusp AI即将完成一轮4亿美元融资,估值达到26亿美元,投资方包括贝佐斯家族办公室和Kleiner Perkins;美国Periodic Labs正推进一轮5亿美元融资,估值约75亿美元;另一家AI科学公司Lila Sciences则计划以85亿美元估值融资20亿美元,潜在投资者包括英伟达旗下NVentures。

随着全球资金持续涌入,中国AI4M赛道也开始迎来属于自己的爆发时刻。近日,深度原理宣布完成A系列融资,累计融资金额近10亿元人民币。其中,孚腾资本领投,华控基金、康君资本(康龙化成CVC)等跟投,顺禧基金、祥峰投资、高瓴创投、戈壁创投、线性资本、联想之星、BV百度风投、启高资本等过半数老股东超额加注。

我们了解到,深度原理不仅拿下了国内AI4M领域迄今最大规模融资,也是行业内少数获得国资重仓支持的AI4M企业。

上海、北京国资联手押注

深度原理此次近10亿元A系列融资,并不是一轮简单的财务融资。

颇为引人注目的是,上海、北京两地国资平台同时出现在投资人名单中。领投方孚腾资本是由上海国投公司作为主要发起人,联合临港集团、上汽集团、宁德时代、哔哩哔哩等领先的产业集团和投资机构共同设立的市场化、专业化股权基金管理人。另一投资方顺禧基金则是北京国管旗下市场化运作的创投基金投资平台,也是北京国管市场化基金群体系的重要成员之一。

对于仍处于早期发展的AI for Science赛道而言,两大地方国资平台联手下注并不多见,也释放出一个信号:AI驱动科学创新,正从资本市场的前沿探索,逐渐进入国家战略布局的视野。

2025年,国务院印发《关于深入实施“人工智能+”行动的意见》,将人工智能赋能科技创新列为培育新质生产力的重要方向。中国科学院院士、鄂维南曾撰文指出,AI for Science的意义不仅在于提升科研效率,更在于推动科研组织方式从“作坊模式”向“平台模式”转变,进而重塑整个创新体系和产业生态。对于中国而言,这场由AI驱动的科研范式变革,也被视为实现2035年建成世界科技强国目标的重要战略机遇。

与此同时,康龙化成旗下康君资本等产业资本的加入,意味着下游产业开始为AI材料研发的商业价值投票。祥峰投资、高瓴创投、戈壁创投、线性资本、联想之星、BV百度风投等过半数老股东选择超额加注(Super Pro-rata),则体现出市场化机构对公司长期发展的持续看好。

“国资+产业资本+市场化VC”集体押注的投资组合,在当下的AI for Science领域并不常见。原因在于,AI for Science至今仍是一个相对“非共识”的赛道。

大模型、AI Agent、具身智能频繁出现在公众视野不同,AI for Science更多活跃于实验室、科研机构和产业研发部门之间。它既缺少面向消费者的爆款应用,也难以用短期收入证明价值,其商业化周期往往以年为单位计算。因此,相较于互联网时代的应用创新,这更像是一场围绕基础科研能力和产业创新效率的长期投资。

2024年,英伟达创始人黄仁勋曾将大语言模型、具身智能和AI for Science并列为人工智能发展的三大方向。在前两者已经成为全球资本竞逐焦点的背景下,AI for Science仍处于早期阶段,也因此成为少数具备巨大想象空间、却尚未形成广泛市场共识的领域。

如今,全球资本正在重新定价这一赛道。以深度原理所在的AI+材料学为例,英国AI材料公司CuspAI即将完成4亿美元融资,投后估值达到26亿美元,较上一轮估值不到一年增长约4倍,投资方包括贝佐斯家族办公室Bezos Expeditions和Kleiner Perkins;由前OpenAI、DeepMind研究人员创立的Periodic Labs,成立不足一年,最新融资估值已升至约70亿美元,而上一轮种子轮估值仅为13亿美元;美国AI科研公司Lila Sciences同样完成数亿美元融资,持续获得全球头部资本支持。

除了在新能源、半导体显示、化工、合金材料等新材料领域具备巨大潜力,以AlphaFold为代表的AI for Science技术在蛋白质折叠、气象预测等特定领域也已经取得了里程碑式成就。

麻省理工博士带队,多项模型实现全面SOTA

和不少AI创业公司一样,深度原理也拥有一支明星团队。

公司由两位麻省理工学院(MIT)博士创立。创始人兼CEO贾皓钧曾在在陶氏化学核心研发部门从事新材料与催化剂研究,希望借助人工智能重新定义材料研发流程;联合创始人兼CTO段辰儒曾任Azure Quantum研究科学家,长期从事量子计算、分子生成模型等方向研究。团队成员则来自MIT、斯坦福大学、浙江大学、上海交通大学等高校,以及微软、谷歌、巴斯夫等企业。

但在AI for Science领域,团队背景只是起点,真正决定竞争力的是能否把算法变成科研生产力。

过去几年,AI已经能够生成分子、预测材料性质,也能完成部分化学反应模拟,但这些能力大多停留在单点工具层面:有的负责生成材料,有的负责预测性质,有的负责实验设计,却很少有公司能够把整个研发流程串联起来。

深度原理选择了一条更难的路径。围绕材料研发,公司自研了Reactive AI平台,并构建起覆盖材料生成、性质预测、化学反应生成等多个环节的模型体系。其中,材料生成模型SAGA能够完成复杂约束下的多目标优化;OA-ReactDiff、React-OT等模型将过去需要数天甚至数月完成的过渡态计算压缩至0.4秒;物性预测模型MPA则在近40项公开任务中达到业内领先(SOTA)水平。

不过,在贾皓钧看来,AI for Science最大的挑战并不是模型,而是如何让模型真正作用于现实世界。科研长期存在"干实验"与"湿实验"脱节的问题:AI负责计算和预测,实验室负责验证,两者之间往往依赖大量人工沟通和反复试错,模型也很难从实验结果中持续学习。为了解决这一问题,深度原理建立了AI Materials Factory高通量实验平台,让AI Agent能够直接控制实验设备完成实验,并自动采集实验数据,再反馈给模型持续迭代,形成"AI预测—实验验证—知识沉淀—模型优化"的完整闭环。

在这一基础上,公司进一步推出AI Scientist平台Mira,将生成式模型、计算工具、客户私有数据库以及实验平台统一连接起来,能够自主完成科研任务拆解、实验设计、执行、复盘和持续优化,实现材料研发流程的自动化。这也是深度原理与不少AI for Materials公司的区别所在。目前,行业里不少企业仍聚焦于解决研发中的某一个环节,例如材料生成、性质预测或实验自动化,而深度原理则试图覆盖从科学假设提出、材料设计、计算验证、实验执行到知识沉淀的完整研发链条,为客户提供一整套材料研发解决方案,而非单一AI工具。

技术之外,更大的考验来自商业化。AI for Science一直被认为是"最难落地"的AI赛道之一。研发周期长、验证成本高、产业需求复杂,使不少创业公司停留在技术展示阶段,真正实现规模化收入的企业并不多。

贾皓钧认为,AI for Science最大的难点很多时候并不是技术,而是产业理解。"很多项目失败,并不是模型不够先进,而是从一开始就没有找准真正需要解决的问题。团队如果全是科学家,却缺少真正理解产业的人,很容易做出技术上漂亮、商业上却没人买单的产品。"

因此,公司从创业之初便围绕产业需求设计研发方向,而不是先做模型,再寻找应用场景。目前,深度原理已服务多家国际及国内头部企业,并实现持续复购。合作模式也从解决单点研发问题,逐步升级为AI Agent订阅以及材料研发管线共研,通过里程碑付款和成果分成等方式建立更深层次合作。业务领域则持续拓展至新能源、电子材料、精细化工、营养日化等多个行业,并已落地多个单客户千万级项目。

注:文/韦香惠,文章来源:投中网(公众号ID:China-Venture),本文为作者独立观点,不代表亿邦动力立场。

文章来源:投中网

广告
微信
朋友圈

FAQ回顾

国内AI4M领域有代表性的创业企业吗?

国内AI4M赛道的代表性企业有深度原理,其近日完成近10亿元A轮融资,是国内AI4M领域迄今最大规模融资,也是少数获得国资重仓支持的AI4M企业。该公司技术能力领先,已服务多家国内外头部企业,落地多个单客户千万级项目。

人工智能发展的三大核心方向是什么?

英伟达创始人黄仁勋曾将大语言模型、具身智能和AI for Science并列为人工智能发展的三大核心方向。其中AI for Science仍处于早期发展阶段,AI4M(AI赋能材料研发)是该领域资本重点布局的热门细分赛道。

深度原理的核心业务能力是什么?

深度原理自研覆盖材料生成、性质预测等环节的Reactive AI平台、高通量实验平台及AI Scientist平台Mira,可覆盖材料研发从假设提出到知识沉淀的全链条,为客户提供一整套材料研发解决方案,而非单点AI工具。

AI for Science赛道商业化落地的难点有哪些?

AI for Science赛道商业化难点主要包括研发周期长、验证成本高、产业需求复杂,不少企业易陷入技术导向而非产业需求导向的误区,若缺少对产业的深度理解,很容易做出技术领先但无商业价值的产品,当前该赛道实现规模化收入的企业较少。

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