Core Insights - World Labs, founded by Fei-Fei Li, has completed a new financing round of $1 billion, bringing its valuation to $5 billion within just 16 months of establishment [1][2][4]. - The company has attracted significant investments from major players like NVIDIA and AMD, as well as new investors including Autodesk, which invested $200 million [3][4]. - World Labs aims to focus on robotics and scientific discovery, actively recruiting engineers to develop core technologies for autonomous navigation and 3D reconstruction [4][5]. Company Overview - World Labs was founded by Fei-Fei Li, a prominent figure in AI, who previously initiated the ImageNet project that significantly advanced deep learning in computer vision [5][6]. - The company has raised a total of approximately $2.3 billion in funding since its inception, with its latest round pushing its valuation to $5 billion [4][6]. - The founding team includes notable scientists and alumni from prestigious institutions, reflecting a strong academic and professional background [6][8]. Product and Technology - World Labs launched its first commercial product, Marble, which generates persistent 3D worlds from various input formats using advanced technologies like Neural Radiance Fields (NeRF) [4][5]. - The business model is based on a freemium approach, allowing users to generate content for free up to four times, with paid plans available for $95 per month [4]. - The company is set to collaborate with Autodesk to enhance its research and modeling capabilities, aiming to build physical AI that enhances human creativity [5][8]. Industry Context - The AI sector is experiencing a wealth creation surge, with companies like Anthropic and Databricks achieving valuations of $380 billion and over $134 billion, respectively [9][11]. - In China, AI companies like Zhiyuan and MiniMax have seen their market capitalizations soar, reflecting the global trend of rising valuations in the AI industry [11]. - The current investment landscape prioritizes "AI content" as a key evaluation criterion for projects, indicating a strong market focus on AI-driven innovations [11].

Core Viewpoint - The Chinese robotics industry is approaching a critical point, with significant market potential highlighted by the recent performance of robots during the 2026 Spring Festival Gala, showcasing advancements in technology and audience engagement [3][4]. Group 1: Industry Developments - The 2026 Spring Festival Gala featured a collaboration of four domestic robotics companies, marking a shift from individual showcases to a collective display of capabilities [7]. - The performance of robots, such as the self-standing action of Yushu Technology's robot, demonstrated advanced dynamic balance and captivated the audience, indicating a growing acceptance of robotics in entertainment [7][8]. - The immediate impact of the gala was evident in e-commerce, with significant sales spikes for featured robots, including a rapid sell-out of products from brands like Magic Atom and Yushu Technology [13][14]. Group 2: Market Dynamics - The Spring Festival Gala has been dubbed a "super roadshow" for robotics companies, with many preparing for IPOs, indicating a potential wave of public offerings in the near future [19]. - Despite the excitement, analysts caution that humanoid robots are still in their infancy, with limitations in their "brains" hindering commercialization [20]. - Current humanoid robot shipments are projected to grow from over 10,000 units in 2025 to approximately 30,000 in 2026, with expectations of reaching 150,000 units by 2030 and potentially exceeding 1 million by 2035 [21]. Group 3: Technological Challenges - The industry is facing challenges in meeting the practical needs of customers, with existing humanoid robots not fully satisfying market demands [20]. - The application of robots is expanding from entertainment to industrial and commercial services, but the complexity of tasks in service and home environments requires more advanced technology and cost considerations [22]. - Three main technological paths are emerging: VLA (Visual Language Action) models, world models, and layered decision-making, each with its own set of challenges related to performance, cost, and operational efficiency [22][23].

一水 发自 凹非寺 量子位 | 公众号 QbitAI 10亿融资，50亿估值。 这就是李飞飞创业公司 World Labs 最新交出的答卷。 今年1月，这家炙手可热的世界模型创业新星就被传出可能正在进行一轮5亿美元新融资，但最终结果却远超预期—— 不是5亿，而是10亿美金 （约合人民币69亿元） ，而且这一轮AMD、英伟达、富达等都投了。 加上本轮融资，成立于2024年4月的World Labs可谓成长飞速—— 不到两年时间，估值从最初的10亿美金直接翻了5倍 。 就这么说吧，可能最接近World Labs融资速度的Anthropic，也花了25个月才达到50亿美元估值。 这速度，你品，你细品。 不到两年，估值一路飙升 进入2026，谁都知道世界模型很hot，但看到World Labs成长如此之快，人们或许才对这一点有了更多实感。 短短一年多时间，这家公司经历了5倍重估。 虽然这里面不乏AI教母李飞飞带来的光环效应，但不得不说，投资者对世界模型的押注决心，比预想中来得更早、也更猛。 这一点，从World Labs本轮投资阵容也可一窥。 根据官方公告，本轮融资吸引了AMD、Autodesk、Emerson Co ...

前Meta公司AI负责人杨立昆（Yann LeCun）也创立了一家"世界模型"初创公司AMI Labs。杨立昆认 为，目前市面上的机器人缺乏对物理世界的基础模型理解，教会AI 理解物理世界的能力至关重要。 谷歌DeepMind也在积极开发世界模型，该公司的Genie模型可以生成和模拟三维环境。 当地时间2月18日，人工智能先驱之一的华人科学家李飞飞创立的AI初创公司World Labs宣布在新一轮 融资中筹集了10亿美元，用于推动所谓的"世界模型"的研发。 在这轮融资中，软件公司欧特克向World Labs投资了2亿美元，其他投资方还包括英伟达以及AMD。尽 管World Labs没有透露估值，但市场预估该公司的估值约在50亿美元。 去年年底，World Labs推出了首个空间智能产品Marble。该公司将Marble描述为一种可以根据图像或文 本提示生成三维世界的基础模型，通过处理来自周围物理环境的视觉数据，开发出高级推理的能力。 借助这笔新融资，World Labs将把重点放在提升机器人和科学发现等领域的应用能力。李飞飞表示，空 间智能模型在未来可以用于增强现实和虚拟现实或机器人。 World Labs" ...

来源|凤凰网财经《公司研究院》 醉拳的意境，在于踉跄之间自有章法。而中国机器人产业的临界点，恰好在 2026年的除夕夜，悄然到来。 当宇树科技的机器人在《武 BOT》节目中"摔倒"在地，随即又一个鲤鱼打挺站了起来，电视机前的观众们议论纷纷，"是故意的，还是失误？" 对此，宇树 CEO王兴兴给出回应称，"剧情需要，打醉拳的时候有要倒不倒的状态，如果机器人倒了以后再自己站起来，会非常帅，非常符合醉拳的 意境"。 以往机器人上春晚，多是拼稳定性、拼队形变换，这次《武 BOT》却让技术去服务剧情和艺术表达。这一摔，摔出中国机器人的技术底气，也摔出一 个即将爆发的市场。 马年春晚开播两小时，京东机器人搜索量环比增长超 300%，客服问询量增长460%，订单量增长150%。新增订单覆盖全国100多座城市，从一线 城市到大小县域均有分布。 从 "假摔"到"真香"，这背后究竟发生了什么？ 01 从独舞到群像 如果说 2025年春晚是宇树科技的独角戏，那么2026年则是一场真正的团体赛。松延动力、宇树科技、魔法原子、银河通用四家国产机器人企业首次 联袂登台，各自亮出看家本领。 松延动力的机器人在小品《奶奶的最爱》中后空翻稳稳落 ...

作者：是地理空间 AI 公司 Niantic Spatial 创始人兼 CEO，在科技与地图交叉领域深耕 25 年，主导开发 了《宝可梦 GO》、谷歌地图、谷歌地球及 Keyhole 地球查看器等产品。 四年前，元宇宙概念正流行时，我曾发文称它是一场反乌托邦噩梦。我当时主张，与其沉浸虚拟世界， 我们更应多与身边的现实世界联结。 时间会证明元宇宙最终能否成真，但 2024 和 2025 年，全球商务与休闲旅行、线下活动参与度均已超过 疫情前水平，这很明显说明：人类本就不该活在头显设备里。 转眼到今天，有人正试图打造这样一个世界：内容越来越不来自朋友，甚至不来自人类网红，而是直接 由 AI 模型生成 —— 一种经过工程化制造的像素与波形，只为抓住你的注意力，直到下一条广告出 现。 短短几年前，先进 AI 刚崭露头角时，人们对未来充满热情：一个人类繁荣的未来，医疗、材料、制 造、机器人技术的突破性进步将让世界变得更好，提升地球上所有人的生活质量。 我认为现在有理由问一句：我们真的在朝这个方向走吗？我们对 AI 的巨额投资中，有多少会真正用于 改善人类生存状况，又有多少会被用来制造更精巧的娱乐与分心手段？ 为什么这件 ...

Investment Rating - The report maintains a strong buy rating for the media industry, indicating a positive outlook for investment opportunities in the AI video sector [2]. Core Insights - The AI video generation technology is evolving rapidly, transitioning from GAN to DiT architectures, which are crucial for advancing towards AGI. This evolution is expected to significantly enhance the capabilities of AIGC (AI-Generated Content) [3][9]. - The global AI video generation market is projected to reach $296 million by 2026, with a year-on-year growth of 35.16%. The industry is exploring both consumer (C-end) and business (B-end) revenue models, with significant advancements in commercial applications expected in the near future [3][4]. Summary by Sections 1. Video Generation Evolution - Video generation integrates multiple modalities, including text, images, and audio, which enhances its complexity and expressiveness, representing the upper limit of AIGC capabilities [7]. - The technology has progressed from early GAN models to the current DiT architecture, marking a significant turning point in the industry with the introduction of models like OpenAI's Sora [9][25]. 2. Technical Progress - Current AI video generation models can produce short segments that approach professional production quality, with resolutions supporting 1080p and frame rates reaching 30fps. However, challenges remain in generating longer videos and maintaining physical realism [34][36]. - The emergence of world models is anticipated to address existing limitations in video generation, potentially leading to a new phase of technological advancement [33]. 3. Commercialization Progress - The AI video generation market is expanding rapidly, with both consumer and business segments progressing simultaneously. The C-end focuses on subscription models, while the B-end primarily utilizes APIs for applications in advertising and e-commerce [3][4]. - The industry is witnessing a shift towards integrating AI capabilities into film production, with significant projects already generating substantial revenue, such as Utopai's projects totaling approximately $110 million [3][4]. 4. Core Beneficiaries - Key companies benefiting from this trend include technology firms with proprietary algorithms, content providers with extensive asset libraries, and platforms actively integrating AI into marketing strategies [4].

| Rank | Model/User | Score | SR | | --- | --- | --- | --- | | 1 F | GigaBrain-0.1/lyf = | 68.34 | 51.67% | | 8 | Spirit-v1.5/Spirit AI | 67.19 | 51.00% | | 3 K | pi0.5/rc_baseline | 61.84 | 42.67% | | 4 | wall-oss-v0.1/Pushi .. : | 55.30 | 35.33% | | 5 | pi0/rc_baseline | 46.41 | 28.33% | | 6 | pi05_generalist/wyf | 31.27 | 17.67% | | 7 | RDT-1B/zsz | 28.84 | 15.00% | | 8 | cogact/hsk | 21.83 | 11.67% | 具身世界模型新一代原生范式重磅登场！继具身基础模型 GigaBrain-0.1 斩获 RoboChallenge 全球第一后，性能更强大的 GigaBrain-0.5M* 又来了。 作为 依托世界模型实现自我 ...

衡宇 发自 凹非寺 量子位 | 公众号 QbitAI 一个 （暂时） 只做具身大脑的公司，抛出了一个 只有2.4B参数 的具身模型。 目前行业风向标如Physical Intelligence的π 0总计约33亿参数，π 0.6的参数量也约莫在50亿以上。 在一个甚至连硬件形态都还没定型的行业里，2.4B参数到底够不够用？ 这家公司给出的答案是，够用。 而且 足以支撑它实时处理三视角的728x728画面，推理延迟仅60毫秒；配合强化学习机制，它还能在真机上不断试错进化。 这就是具身智能创企原力灵机推出的首个具身原生模型产品DM0。 2.4B的轻量小蛋糕， RTX 5090就能跑。 因为从零训练以及对具身数采有不同于行业的看法等原因，该公司称它为"首个具身原生大模型"。 与模型同时发布的还有开源具身原生框架Dexbotic 2.0，以及具身原生量产工作流 DFOL。 这具身软件三件套背后技术路线的操盘手，是原力灵机合伙人、负责大模型的周而进。 他在AI圈早已名声在外。 周而进现在才33岁，但 这人已经在AI领域出名13年了 —— 早在2013年，深度学习和人工智能还是冷门的时候，大二的旷视实习生周而进就以一作身 ...

Core Insights - The article discusses the launch of WorldArena, a unified evaluation system for embodied world models, developed by leading institutions, aiming to shift the focus from visual quality to functional reliability in robotics [1][4][8]. Evaluation Framework - WorldArena introduces a six-dimensional visual assessment framework that includes visual quality, action quality, content consistency, physical adherence, 3D accuracy, and controllability, emphasizing the importance of physical understanding for robots [5][21][25]. - The system also incorporates three embodied tasks to evaluate whether models can effectively participate in real-world tasks, revealing that many visually high-scoring models perform poorly in practical applications [5][27]. EWMScore - EWMScore is a comprehensive scoring system that consolidates various evaluation metrics into a single score, showing a high correlation with human subjective assessments, thus providing a more accurate reflection of model capabilities [6][30][41]. - The correlation between EWMScore and task performance indicates that visual realism does not equate to functional reliability, highlighting a significant gap between visual generation and task execution capabilities [32][44]. Challenges and Future Directions - The article emphasizes that while world models have made significant strides in visual generation, they still face fundamental shortcomings in supporting embodied intelligence tasks and long-term decision-making [33][40]. - The conclusion stresses the need for world models to understand physical laws and maintain consistency in complex environments to transition from being mere visual models to functional embodied intelligence systems [41][45].