卫诗婕 × 郑俊聪

祥峰中国四年不投DeepSeek、智谱、Kimi等大模型，转向AI for science与物理AI，新基金至少50%投具身智能。宇树以成本降一个梯级打开应用场景，量产一致性成为壁垒；中国供应链按周迭代与香港、内地上市，重塑Long China的创新与退出逻辑。

李翔 × 王华东

具身智能正经历中国投资史上前所未有的高估值融资潮，但王华东认为“贵不是问题”，关键在于能否交付有价值、产生毛利的产品。宇树在2021年B轮投前估值约7亿元时，已按出货量位居四足机器人全球第一；行业仍须警惕以更贵、更不可控的方案重做Rule-based已解决的问题，并观察Room-based落地与迭代能否兑现商业化。

Koji × 黄一

22岁的黄一约一年半完成五轮、合计超过一亿美元融资，把RP One发布和量产设为A轮milestone，商业闭环留到B轮。萝博派对以开源本体卡位模型公司与硬件厂商的missing layer，RP0月出货约100台；RP One预计第4季度发布、明年约2000台，但行业到2030年或仅10到15家公司走出。

José × Andrew Kang

Humanoid robotics may be approaching the pre-ChatGPT phase of AI, with $50,000 robots requiring only 20,000 units for $1B revenue and a projected tens-of-billions industry within two to four years.Robo Strategy (NASDAQ: BOT) seeks to challenge SoftBank through public markets, favoring vertically integrated players with embodiment-specific data; US restrictions on Chinese robots could help, while robotics data businesses face a two-to-five-year durability test.

姚颂 × 程曼祺

姚颂判断，VLA的规模化增益正在见顶，物理AI的突破不只取决于模型，而是一场由数据、算力、组织、商业场景与垂直整合共同推进的系统工程。谈过技术路线、创业取舍与泡沫生存法则，他也解释了正行创新为何从轮式双臂和全栈自研出发，押注物理世界的长期价值。

李翔 × 李丰 × 刘鹏琦 × 颜黔杭

上半年一级市场由FOMO、分化和估值前置主导，OpenClaw把agent推向“雇数字员工”。任务轨迹与harness engineering可能在agent层沉淀数据飞轮，Anthropic押注coding，传闻ARR达160亿美元、年终可能盈亏平衡。下半年看真实数据闭环与订单鲁棒性；上市潮若失守预期，或触发一二级倒挂后的“大冷静期”。

卫诗婕 × 赵越

仙工的差异化信号，是覆盖两千多种机型、四百多种传感器并拥有数万台装机的控制器，掌握真实、对齐且持续产生的机器数据入口。赵越更看好先用叉车等垂类专家验证To B的ROI，而非押注全能人形大脑；若端侧模型能理解协议并生成控制能力，现有工具链优势仍可能被抹平。

Avi Felman × Andrew Kang × Jonah Van Bourg

Andrew Kang sees private robotics as crypto in 2014–2015, with market capitalization rising from $20B to $100–150B and potentially 100x upside toward a tens-of-trillions destination.Robo Strategy applies MicroStrategy’s accretive issuance model to private robotics, while a roughly one-year path from GPT-3 to GPT-5-level robotics, rising VC interest and severe failure risk make execution, NAV premiums and valuation marks the key watchpoints.

Peter Diamandis × Salim Ismail × Dave Blundin × Alexander Wissner-Gross

Musk’s Terafab targets 1 terawatt of annual AI-compute capacity versus roughly 20 gigawatts today, making compute the proposed organizing resource of the economy.Demand may absorb every process node as autonomy, AI labor compression, and model distillation expand usage, but ASML, materials, power, labor, capital, and Musk’s timing remain critical execution risks.

卫诗婕 × 宁慕楠 × 姜哲源

机器人春晚的高价赞助与智元自办“机器人奇妙夜”，暴露品牌曝光与工业订单之间的投入产出取舍。2026年行业从Demo转向量产和场景争夺，松延以9998元小布米切入K-12，宇树则保持运控领先。Seedance 2.0能否成为具身训练的世界模型仍无共识，而台上依赖预录与人工遥控，量产、ROI和安全性仍是洗牌关键。