当前，具身智能正面临高质量训练数据严重不足的发展困境。传统的动作捕捉与遥操作技术采集数据， 虽能实现1：1动作复刻，但成本高昂——单台数采设备从几万到十几万不等，且建设大型数采中心投入 巨大。 更关键的是，传统方式采集的数据通用性差，单一设备采集的数据通常仅服务于该设备自身， 形成"数据孤岛"，难以在不同机器人本体间迁移复用。 面对这一困境，行业头部企业纷纷转向互联网视频这一潜力数据源。今年5月，特斯拉工程主管Milan Kovac公开表示，Optimus将告别传统动作捕捉和远程操控，转向直接从互联网视频中学习。互联网视 频资源丰富、成本低廉，且蕴含着人类与物理世界交互的高维信息。如何将这些非结构化的2D视频数 据，转化为机器人可理解、可执行的训练数据，成为技术突破的关键。 枢途科技在成立之初就聚焦在视频提取具身数据技术上，自主研发的SynaData数据管线，实现了从互联 网单目视频中提取多模态、高精度具身训练数据的技术突破。SynaData通过视频数据升维、跨域重定向 retargeting等技术，将普通2D视频转化为包含轨迹、物体网格等信息的具身训练数据。相比传统方式， SynaData成功将综合采集成本 ...

Core Insights - World Labs, co-founded by Fei-Fei Li, has launched a new real-time generative world model called RTFM (Real-Time Frame Model) which utilizes large-scale video data for efficient end-to-end training [3][4] - RTFM can generate new 2D images from one or more 2D inputs without relying on explicit 3D representations, marking a significant advancement in AI rendering capabilities [3][4] - The model can render persistent and 3D-consistent scenes in real-time using a single NVIDIA H100 GPU, enabling interactive experiences in both real and virtual environments [4][10] Company Overview - World Labs was founded in March 2023 by Fei-Fei Li and three other scholars, focusing on developing efficient, scalable, and persistent world models [8][10] - The company raised $230 million in September 2023, achieving a valuation of $1 billion within three months of its establishment [10] - The team consists of approximately 24 members, with a significant representation of Chinese individuals [10] Technology and Innovation - RTFM addresses scalability issues that have long plagued world models, enhancing spatial intelligence in machines, which allows for better navigation and decision-making in complex 3D environments [6][7] - The model's efficiency is highlighted by its ability to support interactive frame rate inference with a single H100 GPU, while its scalability allows for continuous optimization as data and computational power grow [8][10] - Future plans include developing a large model (LWM) that comprehensively understands three-dimensional, physical, and temporal concepts, with applications in AR and robotics [10][12] Research and Development - Fei-Fei Li is also spearheading the Behavior 1K challenge, aimed at standardizing tasks in embodied intelligence and robotics research, providing a platform for training and evaluation [11][12] - The Behavior 1K challenge includes 1,000 tasks focused on long-horizon tasks in everyday environments, promoting collaboration and comparison among researchers [12] - The integration of various AI technologies is seen as a transformative moment for society, emphasizing a human-centered approach in AI development [12][13]

如果说ChatGPT掀起的AI浪潮从消费端向产业端渗透时，ToB领域各个行业与业务单元都步入到了变革 的周期，企业财务领域也位列其中。但仔细回顾可以发现，早期财务领域并未实现大规模企业应用，直 到最近一年随着成本下降，财务领域的AI应用才真正开始落地。过去依赖人工填报、审核、对账的费 控报销流程，正在被AI重新定义。 但另一方面，财务领域作为以 "严谨性""确定性" 为核心诉求的赛道，目前大模型也好，智能体也罢， 不可避免的"幻觉"和可解释性的问题，也是阻碍AI+财务的一大掣肘。近日，全球四大会计师事务所之 一的德勤用AI生成的报告被曝引用虚构文献备受关注。 机遇和挑战并存，或许正代表AI在财务领域的应用走深，处于真正"智能决策"的前夜。 成本下降，带来需求涌现 其实财务真正拥抱AI，并开始从"流程自动化"走向"智能决策"也仅仅才开始不到一年的时间。如果要找 到一个开始的时间节点的话，可以追溯到年初DeepSeek问世之时。 在合思创始人兼 CEO 马春荃告诉笔者，在DeepSeek问世之前，合思就尝试了利用AI大模型的能力去打 造新的财务体系，实现从IT到AI替的过程。但彼时，因为整体模型成本、算力成本很高 ...

智元机器人将批量生产轮式具身机器人。 10月16日上午，智元机器人发布新一代工业级交互式具身作业机器人智元精灵G2。 据悉，精灵G2以工业标准打造，重达185kg，全身拥有26个主动总自由度，其中单手臂为7个自由度， 腿腰部自由度为5个，可实现近似人类的弯腰、转腰和侧向摆动身体能力，并搭载高性能运动关节、高 精度力矩传感器，集成先进的空间感知系统，支持多模态语音交互能力，具备工业、物流、导览等多场 景通用能力。 智元机器人方面表示，目前，精灵G2还未上市就已"售罄"，获得了龙旗科技、均胜集团两个数亿元级 别订单，并已开启首批商用交付。 据笔者了解，今天下午，智元机器人与均普智能（600699.SH）在宁波的G2全球首发暨投产仪式上，智 元向均胜电子交付首批商用的20台精灵G2机器人，并将落地汽车零部件制造场景。 智元机器人合伙人、高级副总裁、具身业务部总裁姚卯青对笔者表示，今年智元机器人交付达到数千 台，相比去年有一个数量级的提升，客户订单包含科研教育、交互服务、工业等领域应用场景。同时， 智元对于机器人系列规划的生命周期约三年。 对于翻跟头等运动控制与交互能力的优势性方面，姚卯青进一步表示，人形机器人最重要 ...

Core Insights - The annual e-commerce event "Double 11" has commenced, with significant consumer engagement and sales growth reported by JD.com [2][3] - JD.com has shifted its strategy this year by eliminating pre-sales and focusing on immediate sales at the lowest prices during the event [2] - The company is emphasizing a straightforward theme of "good quality and low prices" to attract consumers [3] Sales Performance - JD.com reported a 47.6% year-on-year increase in active users on its app from October 9 to 10, leading the industry [2] - Orders for major categories such as home appliances, mobile phones, and digital products saw over 70% year-on-year growth by October 14 [2] Marketing Strategy - JD.com is enhancing its marketing approach by introducing "Surprise Days," which focus on specific categories and consumer interests, starting from October 16 [3][4] - The company is leveraging its procurement team to create a "Procurement Recommendations" section, showcasing high-quality products at competitive prices [5] Product Offerings - JD.com is set to launch a "National Good Car" in collaboration with CATL and GAC Group during the Double 11 event, available exclusively on its platform [5] - The company is also promoting its local lifestyle services, partnering with over 3 million merchants to provide a wide range of offerings [6] Logistics and Technology - JD.com is implementing advanced logistics solutions, including the "Super Brain + Wolf Pack" intelligent device cluster, to enhance operational efficiency during the event [8] - The integration of AI and robotics in logistics is expected to improve service quality and reduce costs for brand partners [8][9] Conclusion - The Double 11 event continues to evolve, with JD.com focusing on consumer needs and leveraging technology to enhance the shopping experience [9]

Core Insights - Polestar has closed its only direct sales store in Shanghai, marking a strategic shift in response to the evolving landscape of the electric vehicle industry and the challenges of the direct retail model [1][4][12] Company Overview - Polestar, originally a performance model series of Volvo, became an independent brand in 2017, targeting the high-end electric vehicle market with backing from Volvo and Geely [1] - The brand initially gained traction with the Polestar 1 hybrid sports car priced at 1.45 million yuan, creating a high-end image [1] Strategic Shifts - Following its initial success, Polestar shifted focus to the mass market, launching the Polestar 2 at a significantly lower price range of 250,000 to 300,000 yuan, and later introducing the Polestar 3 and Polestar 4 at varying price points [2] - This inconsistent pricing strategy has left consumers confused about Polestar's brand identity, failing to establish a luxury aura while lacking competitive pricing in the mass market [2][3] Sales Performance - In 2023, Polestar's global sales reached approximately 54,000 units, falling short of its target of 80,000 units set at the beginning of the year [2] - The brand's performance in China has been particularly poor, with sales dropping from 2,048 units in 2021 to just 1,100 units in 2023 [2] Management Challenges - Polestar has experienced significant management turnover in China, with seven different leaders in eight years, leading to instability and a lack of long-term strategic execution [3] - Key decision-making authority remains centralized in Sweden, hindering Polestar's ability to respond quickly to market changes [3] Cost-Cutting Measures - The closure of the Shanghai store is part of a broader global strategy to cut costs, including plans to lay off approximately 450 employees, which is about 15% of its workforce [4][5] - Polestar is transitioning to a "sales-driven production" model, aligning more closely with Geely's sales and service network [4] Industry Context - The challenges faced by Polestar reflect broader issues within the electric vehicle sector, where many brands struggle to find a sustainable retail model [6][12] - The high costs associated with direct retail, including significant rental and operational expenses, have led many brands to reconsider their strategies [7][9] Retail Model Comparison - The debate between heavy asset direct sales and light asset dealership models continues, with each having distinct advantages and disadvantages [9][10] - Some brands are exploring hybrid models that combine direct sales in major cities with dealership networks in lower-tier markets to balance control and expansion [10][11] Future Outlook - The closure of Polestar's direct store is seen as a new beginning rather than an end, indicating a shift towards more sustainable business models in the competitive electric vehicle market [12]

Core Insights - The global financial markets experienced significant turbulence following the National Day holiday in 2025, with the A-share Shanghai Composite Index dropping by 0.94% and the Nasdaq index falling by 3.56% on October 10. The cryptocurrency market faced even harsher conditions, with a record liquidation amount of $13.475 billion within 24 hours on October 11, marking the highest single-day liquidation in history [1][2]. Group 1: Market Reactions and Causes - The "1011 Black Swan" event was triggered by macroeconomic risks, market structure imbalances, and a crisis of technical trust, rather than being an isolated incident [1]. - The immediate catalyst for the market crash was the announcement by former President Trump on October 10 regarding a 100% tariff on Chinese goods and expanded export controls, which led to a global risk-off sentiment and a downgrade in the World Trade Organization's 2026 global goods trade growth forecast to 0.5%, the lowest since 2009 [1][2]. Group 2: Cryptocurrency Market Dynamics - In 2025, the correlation between cryptocurrencies and traditional financial assets reached a historical peak, with Bitcoin's 30-day correlation with the S&P 500 rising to 0.78, indicating that 40% of cryptocurrency price fluctuations could be explained by S&P 500 volatility [2][6]. - The leverage in the market significantly increased, with retail investors' average leverage ratio soaring to 10 times, and the overall market leverage ratio reaching a high of 38% since May 2022 [2][3]. Group 3: Technical and Structural Issues - The USDe stablecoin's 12% subsidy policy led to a leveraged entry into the market, creating a false sense of prosperity. However, this model's vulnerability became apparent when collateral prices fell, triggering a "death spiral" in the leveraged market as forced liquidations exacerbated price declines [3][5]. - The absence of major market makers during the crash period contributed to a liquidity crisis, as algorithmic trading triggered stop-loss orders at critical support levels, leading to a rapid price decline [3][4]. Group 4: Trust and Security Concerns - The market faced a trust collapse due to technical security anxieties, particularly regarding quantum computing threats. Over 60% of Bitcoin supply is stored in addresses vulnerable to quantum attacks, raising concerns about the security of digital assets [4][5]. - The USDe de-pegging incident, where it fell to $0.62 (a 38% drop), highlighted the risks associated with stablecoin mechanisms and the lack of liquidity during extreme market conditions [5][7]. Group 5: Regulatory and Market Sentiment - The tightening of global regulatory policies was evident, with the SEC and other international bodies initiating actions against unregulated crypto exchanges, leading to a significant drop in trading volumes [8]. - Despite the recent downturn, the underlying drivers of the cryptocurrency bull market, such as global liquidity easing, remain intact, suggesting that the market may be undergoing a deep adjustment rather than a complete reversal [9][10].

特斯联副总裁杨立德（前排右），新华三集团AI与创新科技系统部总监杨洋（前排左）代表双方签 约。特斯联高级副总裁刘斌（后排右三），副总裁唐智（后排右二）、新华三集团副总裁、企业事业部 总经理管志强（后排左三）等见证签约 特斯联与新华三正式宣布达成战略合作，双方将集中优势资源，围绕AIoT算力平台打造及异构算力生 态建设进行深度合作。借助特斯联在异构算力编译与空间智能场景中的业务实践与AIoT技术沉淀，及 新华三在"云-网-安-算-存-端"全产业链的深度布局，双方的合作将进一步推动国产算力在AIoT领域的创 新与应用，加速构建高效、智能、安全的异构算力生态，共同推动异构算力融合的发展和应用。 3. 全新探索：异构算力的超节点建设 当前AIoT场景的算力需求已不再是"单一维度的算力叠加"，而是"多维度算力的精准匹配"，这种需求升 级直接催生了对异构算力超节点的需求。任务类型"多样化"、场景"规模化"、响应时效"实时化"，都对 算力集群的协同提出了更多挑战。而异构算力超节点是构建AIoT算力生态基础设施的"基石"。 新华三构建的覆盖智算领域的全栈技术架构、算力供给、场景落地等技术壁垒，为搭建异构算力超节 点，奠定了坚 ...

Group 1 - ABB Group has sold its robotics business unit to SoftBank for $5.375 billion, with the transaction expected to be completed in mid-2026 pending regulatory approval [1][2] - The sale is seen as a strategic move by ABB to create immediate value for its shareholders, as the robotics division has been underperforming, with a 39% year-on-year profit decline in 2024 [2][3] - SoftBank's acquisition aims to enhance its ambitions in physical AI, which focuses on the application of artificial intelligence in the real world, moving beyond traditional software algorithms [1][6] Group 2 - The robotics market has become increasingly competitive, with ABB's profit margins in this sector dropping to 12.1%, below the overall group profit margin of 18.1% [3][4] - The decline in ABB's robotics business is attributed to reduced orders from key industries like automotive and electronics, as companies delay automation investments due to economic conditions [4][5] - The rise of domestic Chinese brands in the robotics market, which increased their market share from 47% in 2023 to 58% in 2024, has also pressured ABB's pricing strategies [4][5] Group 3 - SoftBank's previous investments in robotics, such as Aldebaran and Boston Dynamics, faced challenges, but the acquisition of ABB's robotics business is seen as a more strategic fit due to its established market presence [6][7] - SoftBank aims to create a comprehensive ecosystem that integrates hardware, AI capabilities, and data feedback loops to develop a self-learning physical intelligence network [7][10] - The acquisition positions SoftBank as a more active player in the robotics industry, transitioning from an "ecosystem investor" to an "ecosystem leader" [10][11]

大家好！来自华盛顿特区的问候。 我是美国全国亚太裔商业与创业协会总裁兼首席执行官董继玲（Chiling Tong），。我们代表着全美超 过 300万名亚太裔小企业主，并与全国约130家商会及合作机构密切协作。非常荣幸能够参与在斯坦福 大学举办的 2025 NEX-T峰会。 首先，我要感谢 NextFin.AI、全球亚裔领袖联盟、钛媒体集团、盛大集团及巴伦中国以及所有杰出领导 者和合作伙伴，他们为这场适逢其时、前瞻性大会付出了努力。 美西时间2025年9月27–28日，钛媒体集团携手NextFin.AI、全球亚裔领袖联盟（GALA）、盛大集团 （Shanda Group） 与 巴伦中国（Barron's China），在美国斯坦福大学成功举办了首届硅谷未来峰会 ——NEX-T Summit2025。 美国国家亚太裔商业与创业协会总裁兼首席执行官董继玲（Chiling Tong）在致辞中表示，人工智能与 全球格局的转变正处于关键拐点。她强调，亚裔美国企业家与创新者应携手推动技术与商业进步，搭建 文化桥梁、增强信任，共同塑造一个更具包容性的全球未来。 以下是董继玲（Chiling Tong）在钛媒体硅谷未来峰会—— ...