RoboStrategy's Kang Bets Humanoid AI Will Disrupt $50 Trillion Labor Market

RoboStrategy的Kang押注人形AI将颠覆50万亿美元劳动力市场 AI推理能力的加速以及每小时低于2美元的运营成本将推动人形机器人在各行业快速普及，取代体力与脑力劳动岗位。 本期节目探讨了人形机器人与具身智能机器人从实验室概念向商业化落地部署的快速转变。Andrew Kang强调，通用智能与实体机器人的融合将使体力劳动实现产品化，大幅降低劳动力成本，并从根本上重塑全球劳动力市场。讨论内容涵盖了诸如全民基本收入（Universal Basic Income）等政策转变的潜在必要性，以及RoboStrategy（BOT）背后的独特投资逻辑——该上市公司旨在为散户投资者提供参与高增长、拟上市机器人公司的流动性投资渠道。 具身智能与机器人市场正迎来关键拐点，2025年全球人形与四足机器人出货量同比增长250%，达到约53,000台便是明证。 (1:23) 高盛将其对人形机器人潜在市场总量（TAM）的预测上调了6倍，预计到2035年TAM将达380亿美元，且到2030年年出货量将突破250,000台。 (1:23) 正如大语言模型将知识工作产品化一样，人形机器人正在将每年达50万亿美元的全球体力劳动市场产品化。 (2:55) 机器人的核心瓶颈已从通用智能和软件推理，转移至物理制造能力和规模化生产基础设施的建设。 (7:36) 人形机器人相比传统人类劳动力具有显著的成本优势，其有效成本估计约为每小时2.00美元，而人类产业工人为每小时35至40美元。 (7:36)

Andrew Kang, CEO of RoboStrategy, argues that the convergence of large language models and embodied physical AI is driving an accelerated transformation of the global labor economy. To capture value before these frontier firms reach public markets, RoboStrategy is using a closed-end public vehicle to invest directly in leading private robotics developers. Economic Drivers of the Physical AI Transition (1:23) The commercialization of humanoid and quadruped robotics is moving from experimental lab demonstrations to active factory infrastructure. Global shipments expanded 250% year-over-year to roughly 53,000 units in 2025, with Goldman Sachs projecting annual shipments to top 250,000 units by 2030 and revising its 2035 addressable market estimate upward by sixfold to $38 billion. Kang attributes this acceleration to general-purpose AI solving spatial reasoning bottlenecks far faster than historical projections anticipated, targeting a global physical labor pool valued at approximately $50 trillion annually. Adoption is primarily propelled by stark labor economics. An effective operational cost of approximately $2.00 per hour for a humanoid unit capable of replacing three shifts compares to $35.00 to $40.00 per hour for typical US industrial human labor, while eliminating benefits like healthcare and retirement matching. Kang highlights unedited factory trials, such as Figure AI operating autonomously for eight consecutive days with battery swaps, to show that software reliability is already functioning. Consequently, the primary operational bottleneck has shifted from intelligence algorithms to factory scale, with Tesla already planning production capacity for 10 million humanoid units annually. Labor Displacement and Emerging Policy Responses (13:59) Kang contends that unlike past industrial shifts where displaced workers migrated into new mechanical or technical trades, modern physical AI targets both physical dexterity and cognitive decision-making concurrently. This simultaneous scaling removes traditional fallback roles, as even secondary positions like robot maintenance will likely be automated by robotic labor over time. Because market-driven automation could systematically erode both cognitive and manual jobs, Kang views Universal Basic Income (UBI) as an inevitable policy measure rather than an optional safety net. Host Josh Brown and Kang note that state responses may ultimately mirror sovereign wealth structures, similar to the Nordic model. The US government is already establishing precedents for direct strategic industrial intervention through the CHIPS Act, domestic rare earth processing subsidies, and emerging federal legislative priorities aimed at preserving national competitiveness in robotics. RoboStrategy's Closed-End Venture Arbitrage Model (21:55) Because mega-cap technology firms remain private substantially longer due to abundant venture capital, public market retail investors are typically excluded from pre-IPO appreciation. Operating under the Nasdaq ticker BOT, RoboStrategy functions as a publicly traded closed-end fund designed to provide liquid exposure to private robotics equity. The vehicle applies an issuance playbook modeled after MicroStrategy's Bitcoin accumulation strategy, executing private placements at premiums to its Net Asset Value (NAV) to acquire underlying private company equity. Over a three-week stretch, RoboStrategy secured roughly $100 million across three Private Investment in Public Equity (PIPE) deals at prices roughly 3x to 4x its publicly stated NAV. Kang asserts that issuing public shares at an active premium allows the vehicle to purchase private shares on an accretive basis, continually seeking to expand the volume of targeted robotics assets held per public share while exploiting multiple-arbitrage spreads between private and public market valuations. Concentrated Portfolio Bets and Mechanical Approaches (31:30) RoboStrategy runs a concentrated portfolio focused on private industry leaders rather than broad sector diversification, adhering to a power-law thesis that outsized venture returns accrue to market pioneers. Its primary holdings include Figure AI, positioned as a vertically integrated premium brand focusing on full humanoids; Apptronik, an established joint-actuator specialist collaborating with Jabil on manufacturing and Google DeepMind on artificial intelligence; and Doosan Robotics, which focuses on non-bipedal, stationary, or wheeled industrial systems. Kang emphasizes that non-bipedal designs address a massive, cost-efficient industrial segment. By eliminating the mechanical complexity and elevated bill of materials associated with robotic legs, stationary dual-arm and wheeled systems can tackle single-plane factory and logistics tasks immediately, sidestepping the manufacturing costs required for fully mobile bipedal humanoids. RoboStrategy's strategy depends heavily on public markets maintaining an equity premium over fund net asset value to fund private venture acquisitions, alongside the successful commercial scaling of hardware manufacturing among its concentrated private bets.

RoboStrategy联合创始人兼首席执行官Andrew Kang认为，随着大语言模型与具身智能加速融合，人形机器人正在将全球体力量化为高性价比的标准化产能；面对技术替代对劳动力市场的颠覆性冲击，投资者与政策制定者不仅需要直面全民基本收入等制度设计，还需借助创新的公开资本工具参与私有前沿机器人企业的增长红利。 物理AI拐点与劳动力成本结构性重塑 (2:55) Kang指出，物理AI领域在2023至2024年大语言模型爆发后迎来了关键转折，具身智能的发展将机器人攻克空间推理能力的时间窗口缩短至三到七年。根据高盛的预测数据，全球人形机器人长期潜在市场空间（TAM）已被上调6倍至2035年的380亿美元，预计到2030年年出货量将突破25万台；而2025年人形与四足机器人的全球出货量预计将同比增长250%至约53,000台。全球体力劳动市场年规模约为50万亿美元，人形机器人的本质是将体力劳动实现产品化交付，从而打破传统人力在工作时长、工伤风险、医疗保险以及401(k)养老金等方面的刚性约束。 在商业落地验证方面，实际产业推进速度显著快于公众认知。Kang以Figure AI在工厂进行的连续8天无人工遥操作物流测试为例，自主机器人每8至10小时自动轮换更换电池组，展现了执行分拣与搬运任务的稳定性。从经济效益对比，美国传统工业工人平均时薪约为35至40美元，而单台人形机器人的折算运营成本仅约为每小时2美元（可替代约3名工人）。Kang认为，行业当前的演进瓶颈已非纯软件算法，而是以特斯拉规划年产1000万台为代表的规模化实体制造与量产供应链建设。 机器替代的双重冲击与政策应对框架 (13:59) 针对行业普遍宣称“机器人仅会增强而非替代人力”的观点，Kang提出了明确反驳。他认为，以往技术革命中“铁匠转行为汽车维修工”的就业过渡模式如今不再成立，因为早期的机械发明缺乏认知推理能力，而当前的物理AI与底层大模型正在同步替代人类的重体力劳动与复杂认知决策，未来甚至机器人的维护与制造也将由机器人自主完成。由于传统以规划与空间推理为退路的就业缓冲带大幅收窄，即便是在依赖人工提供差异化服务的高端酒店餐饮业，未来也可能面临高拟真度智能机器人的竞争。 基于双重替代引发的失业风险，Kang主张必须在危机全面爆发前系统性规划全民基本收入（UBI）等保障体系，单纯提供智力工具并不能直接解决公众的基本生存需求。在探讨财富分配模式时，主持人Josh Brown将其比作通过国家石油财富向全体公民发放红利的北欧主权财富基金模型，Kang对此表示认同，并指出美国政府当前通过《芯片与科学法案》资助半导体制造业、注资稀土加工领域以及国会酝酿中的机器人产业激励政策，反映出国家层面的战略资本正加速介入关键工业技术以巩固竞争力。 幂律分布下的核心标的配置策略 (31:30) 在投资组合管理上，Kang表示物理AI与机器人领域的初创投资将呈现极端的幂律分布特征，资本将高度向具备顶级工程人才与雄厚资金储备的头部领跑者聚集。RoboStrategy目前采取高集中度配置策略，其重点持仓标的涵盖不同技术路线与市场切入点。其中，Figure AI被定位于“人形机器人领域的苹果”，专注于高阶软硬件垂直整合与精致的用户体验；其面向消费级市场的最终定价虽未明确（高端机型或达六位数），但工业端正加速推进租赁运营模式。 与此同时，组合在形态与零部件供应链上进行了针对性差异化配置。拥有近十年研发历史的Apptronik深耕关节致动器（actuator）等核心硬件开发，并依托与代工制造巨头Jabil以及Google DeepMind的战略协同加速量产推进；而斗山机器人（Doosan Robotics）则主打固定双臂工作站与轮式移动底盘等非双足方案。Kang分析认为，双足结构大幅抬高了机械复杂性与制造成本，在绝大多数平坦地面、任务重复且无需爬楼的工厂场景中，非双足方案具备更高的量产可行性与资本回报效率。 封闭式基金架构与一二级市场套利 (21:55) 为了让公开市场投资者能穿透传统风投壁垒，RoboStrategy以公开上市的封闭式基金形式在纳斯达克交易（代码：BOT）。Kang指出，当前以SpaceX、Anthropic和OpenAI为代表的硬科技独角兽依靠充沛的私人资本长期保持非上市状态，传统被动指数基金与公开市场散户无法参与早中期成长红利；而BOT直接持有私有机器人企业的股权资产，为二级市场提供了流动性工具。 在资本运作层面，RoboStrategy复制了MicroStrategy在加密资产上的资本运作思路。针对增发稀释股权的疑虑，Kang解释公司在三周内通过三笔PIPE（私募股权投资于上市公司）交易融资约1亿美元，发行价格较公开披露的基金净资产值（NAV）享有约3至4倍的溢价。由于一级市场资产交易存在价格滞后与折价，而公募市场具有更广泛的全球流动性与重估乘数（类似于企业并购中将非上市标的估值重新锚定至公募高倍数），管理层通过溢价发行股票筹集资金并回购头部私有股权，旨在持续推高每股内含的机器人资产敞口。 物理AI与人形机器人的普及展现了颠覆传统生产力与劳动力结构的确定性趋势，然而该产业的大规模兑现仍受制于实体量产产能爬坡以及AI安全对齐等深层挑战；此外，RoboStrategy等通过公募溢价反哺一级市场资产的资本创新工具，其长期有效性亦面临非公开估值透明度不足与二级市场溢价波动等结构性考验。

## Background 本视频由知名财经频道 The Compound 于 2026 年 7 月 13 日发布在 YouTube 平台。节目由 Downtown Josh Brown 主持，对话嘉宾为 RoboStrategy（Nasdaq：BOT）首席执行官兼联合创始人 Andrew Kang，探讨了人形机器人与具身智能（Physical AI）从实验室走向工业量产的拐点、对就业市场的颠覆性冲击，以及通过公募封闭式载体捕捉私有机器人龙头成长红利的投资逻辑。 ## TL;DR Andrew Kang 阐述了他从早期加密投资转向全力押注人形机器人与具身智能的核心逻辑，指出随着大模型通用认知能力的攻坚，机器人将在未来 3 到 5 年内迎来量产拐点并重塑价值 50 万亿美元的全球物理劳动力市场。面对机器人每小时约 2 美元的极致成本对人类工人的替代，他打破了以往技术升级必将催生新工种的幻想，呼吁提前做好全行业失业应对与 UBI 制度准备。为了解决散户无法参与顶级私有化机器人企业成长的痛点，他创立了上市载体 RoboStrategy（BOT），通过类似 MicroStrategy 的溢价定增与倍数套利飞轮，重仓 Figure AI、Apptronik 和 Dyna Robotics 等前沿领跑者。 ## Atmosphere & style 整场对谈节奏紧凑犀利、充满干货，Josh Brown 标志性的风趣自嘲与接地气追问，与 Andrew Kang 冷静笃定、直言不讳的硬核科技现实主义形成了强烈的化学反应。对话既毫无保留地揭示了未来机器淘汰脑力与体力劳动的残酷社会现实，又充溢着资本前沿掘金万亿超级赛道的敏锐与亢奋。聆听体验既像在近距离窥探硅谷实验室里近乎魔幻的未来图景，又让人深切感受到站在历史产业剧变十字路口上的危机感与探索欲。 ## Bulletin - [01:23](https://www.youtube.com/watch?v=1Tgw82x46YA&t=83s) 全球人形与四足机器人出货量去年增长250%至近53,000台，高盛预计2035年市场规模将达380亿美元，较此前预测上调六倍。 - [05:16](https://www.youtube.com/watch?v=1Tgw82x46YA&t=316s) 具身智能瞄准每年约50万亿美元的全球物理劳动力市场，人形机器人综合使用成本约2美元/小时，远低于美国蓝领工人35至40美元的时薪水平。 - [12:08](https://www.youtube.com/watch?v=1Tgw82x46YA&t=728s) 行业认知智能预计在未来1至2年内成熟，核心瓶颈已转向制造业量产产能爬坡，特斯拉正规划年产1,000万台规模的超级工厂领跑赛道。 - [21:56](https://www.youtube.com/watch?v=1Tgw82x46YA&t=1316s) RoboStrategy（Nasdaq: BOT）作为公开上市的封闭式基金，通过复刻 MicroStrategy 溢价增发机制投资 Figure AI、Apptronik 等未上市龙头，为散户提供流动性桥梁。 ## Key takeaways - [02:55](https://www.youtube.com/watch?v=1Tgw82x46YA&t=175s) 大语言模型的爆发将通用智能的攻克周期大幅缩短至3到7年内，使人形机器人从长达数十年的不可行空想转变为具备极高商业确定性的物理AI投资赛道。 - [05:16](https://www.youtube.com/watch?v=1Tgw82x46YA&t=316s) 人形机器人的本质是将物理劳动力产品化，其覆盖空间达到全球每年50万亿美元的体力劳动市场，并享有24小时轮班、零停工与零社保成本等压倒性优势。 - [10:36](https://www.youtube.com/watch?v=1Tgw82x46YA&t=636s) 人形机器人执行包裹分拣并实现连续8天自主运行被视为具身智能的“Mini ChatGPT时刻”，证实多机轮班换电自主作业已在工业场景进入实用阶段。 - [12:08](https://www.youtube.com/watch?v=1Tgw82x46YA&t=728s) 机器人普及的技术阻碍已基本消除，当前核心瓶颈在于物理层面的超级工厂量产与产能爬坡，预计将在未来3至4年内逐步打破。 - [14:00](https://www.youtube.com/watch?v=1Tgw82x46YA&t=840s) 与以往技术革命仅替代基础体力不同，物理AI与具身智能同时攻克复杂认知规划与通用思维，“技术升级自然催生更高阶人类就业”的历史外推规律本次将彻底失效。 - [14:00](https://www.youtube.com/watch?v=1Tgw82x46YA&t=840s) 连机器人制造与维修等新工种未来都将由机器人自主执行，面对结构性劳动力替代，全民基本收入（UBI）将成为政府不可回避的社会生存底线。 - [18:02](https://www.youtube.com/watch?v=1Tgw82x46YA&t=1082s) 若AI安全与对齐问题妥善解决，劳动力成本归零将带来社会生产力的指数级爆发，但必须辅以类似主权财富基金的再分配机制以防财富高度极化。 - [21:56](https://www.youtube.com/watch?v=1Tgw82x46YA&t=1316s) RoboStrategy 设立为上市交易封闭式基金（Nasdaq: BOT），旨在打破顶级AI硬件初创长期不上市的壁垒，让公开市场投资者能参与未上市领头羊的高倍成长。 - [23:17](https://www.youtube.com/watch?v=1Tgw82x46YA&t=1397s) 传统资产净值（NAV）与私募二级市场存在流动性折价和估值滞后，无法匹配物理AI呈指数级突飞猛进的估值迭代，导致公募股票常年以高溢价交易。 - [31:30](https://www.youtube.com/watch?v=1Tgw82x46YA&t=1890s) 基金投资遵循风投的幂次法则，坚决回避平庸的低估值追随者，重仓 Figure AI、Apptronik 和 Dyna Robotics 等占据技术壁垒与顶级团队的赛道赢家。 - [33:42](https://www.youtube.com/watch?v=1Tgw82x46YA&t=2022s) Figure AI 走全栈自研的极致产品化路线，Apptronik 凭借执行器核心元老积淀与谷歌 DeepMind 及捷普深度合作，Dyna Robotics 则专攻高性价比的非双足轮式工业工位。 - [36:51](https://www.youtube.com/watch?v=1Tgw82x46YA&t=2211s) RoboStrategy 复刻 MicroStrategy 模式，以远高于NAV的溢价向长线机构完成近1亿美元定增，通过高效配置持续增厚每股内含的优质机器人企业股权。 - [39:44](https://www.youtube.com/watch?v=1Tgw82x46YA&t=2384s) 由于一级市场资金极其充沛、IPO审核严苛，硬科技独角兽保持私有化的时间大幅延长，传统机器人ETF已无法精准捕获前沿具身智能的核心爆发期。 ## Detailed walkthrough ### [00:17](https://www.youtube.com/watch?v=1Tgw82x46YA&t=17s) 赞助商信息与节目开场 本期节目由 Betterment Advisor Solutions 赞助播出。对于投资顾问公司（RIA）而言，普遍面临一个矛盾：既不希望拒绝新客户或设定过高门槛，希望陪伴处于起步阶段的客户共同成长以建立长期合作关系；但另一方面，这些体量较小的基础账户往往需要耗费大量后台人力进行开户、交易执行与动态再平衡，令运营团队不堪重负。Betterment Advisor Solutions 通过无纸化全自动的入驻流程、精简且兼顾税务效率的投资组合管理，协助投资顾问实现客户分层管理，在降低运营压力的同时维持高标准服务。 ### [01:23](https://www.youtube.com/watch?v=1Tgw82x46YA&t=83s) 具身智能崛起与“RoboStrategy”上市 Downtown Josh Brown 正式引出本期核心主题：人形机器人已不再是科幻小说中的概念，而是正在实打实地走进工厂车间，未来数年将迎来从早期采用者到大规模普及的爆发拐点。去年全球人形机器人与四足机器人的出货量激增 250%，达到近 53,000 台。高盛（Goldman Sachs）预测，到 2030 年人形机器人年出货量将突破 25 万台；到 2035 年，潜在市场总额（TAM）将达到 380 亿美元——该数字较其此前的预测上调了整整六倍。 本期嘉宾 Andrew Kang 便是极少数早早洞察这一趋势的投资者。作为加密投资领域的资深人士，他曾参与早期去中心化金融（DeFi）的投资浪潮，随后毅然转身全力押注“物理 AI”（Physical AI / 具身智能）。在全行业风投对其投资意向普遍持否定态度的情况下，他不仅自掏腰包下注，还在纳斯达克成立了公开上市的投资工具 RoboStrategy（股票代码：BOT），使散户投资者也能直接参与机器人革命的浪潮。 - 去年全球人形与四足机器人出货量：近 53,000 台（同比增长 250%） - 高盛预测 2030 年人形机器人年出货量：超 250,000 台 - 高盛预测 2035 年人形机器人潜在市场规模（TAM）：380 亿美元（最新上调 6 倍） - 嘉宾背景：Andrew Kang，RoboStrategy 联合创始人兼 CEO（Nasdaq 代码：BOT），前加密/DeFi 早期资深投资人 ### [02:55](https://www.youtube.com/watch?v=1Tgw82x46YA&t=175s) “买下该死的机器人”：从科技股长牛到物理 AI 的转向 Josh 回忆道，早在 2017 年，他曾在当时风靡的金融博客圈撰写过一篇题为《买下该死的机器人》（"Just Own the Damn Robots"）的文章。当时他指出，纳斯达克爆发的核心逻辑在于市场开始意识到科技巨头正站在全面自动化经济的前沿，对于担忧未来的普通投资者而言，最佳策略就是做多科技股。此后纳斯达克 100 指数大涨约 500%，证明了当年的判断。然而，Andrew Kang 走得更远，不仅是配置科技股，而是将个人与事业全部押注在实体机器人上。 谈及转变的契机，Andrew 透露大约两三年前（2023 年底至 2024 年初），一位朋友向他展示了 Figure AI 的展示视频。画面中，人形机器人在实验室里像真正的人类一样自如行走，极具科幻电影的质感。放在过去，人们会觉得这种设想遥不可及；但在 ChatGPT 发布后的大背景下，Andrew 意识到底层逻辑彻底变了。过去人形机器人之所以是“空想”，关键瓶颈始终在于机器人缺乏通用智能；而随着大语言模型与人工智能进化的急剧加速，在 3 到 5 年至 7 年的合理周期内解决机器人的通用智能已变得完全可行，这彻底点燃了他的投资决心。 ```mermaid flowchart TD accTitle: 具身智能投资逻辑演进 accDescr: 展示从传统认知瓶颈到大语言模型突破后机器人产业具备可行性的逻辑路径 subgraph Bottleneck["过去数十年的行业瓶颈"] A["硬件与外观初具雏形"] --> B["缺乏通用智能大脑"] B --> C["被视作 50-100 年的不可行空想"] end subgraph Breakthrough["大模型突破带来转折"] D["ChatGPT 与生成式 AI 爆发"] --> E["通用智能研发周期大幅缩短"] E --> F["预期 3-7 年解决机器人智能瓶颈"] end C -.->|"技术范式转变"| D F --> G["具身智能具备商业与投资可行性"] ``` ### [05:16](https://www.youtube.com/watch?v=1Tgw82x46YA&t=316s) 逆风投共识押注 Figure AI：50 万亿美元物理劳动力市场 在确定投资意向前，Andrew 征询了身边风投圈的大量人脉，几乎所有人都劝他不要投资 Figure AI。然而他最终依然选择力排众议，先是自掏腰包投入 100 万美元，随后追加投资将持仓扩大到了 1,900 万美元。谈及当时为何敢于逆行，Andrew 解释称，当时必须建立一套全新的分析框架，因为此前市场上根本不存在人形机器人这一成熟品类，没有任何历史数据可以直接套用。 Andrew 的核心逻辑在于：大语言模型（LLM）将知识工作与数字智能产品化，而人形机器人则是将“物理劳动力”（Physical Labor）产品化。只要机器人能够胜任人类所能从事的任何物理工作，它就拥有无限优势——它可以 24 小时全年无休运转、无需睡眠、不知疲倦、从不抱怨也不会离职，更不需要企业缴纳健康保险或提供 401(k) 退休金配比。从这个维度切入，其潜在市场空间是全球整个物理劳动力市场，其年经济价值高达约 50 万亿美元。不仅如此，50 万亿美元只是当前在岗物理劳动力的价值，尚未计入目前无法招募到人手的劳动力缺口，以及未来机器人在太空或其他人类无法涉足的极端环境中所开辟出的全新增量市场。 - 初始投入与加仓规模：从最初的 100 万美元追加至 1,900 万美元 - 全球物理劳动力市场年价值基准：约 50 万亿美元 - 人形机器人相比人类劳工的核心优势：全年无休、无疲劳停工、无离职风险、零医保与养老金成本 ### [07:36](https://www.youtube.com/watch?v=1Tgw82x46YA&t=456s) 采纳周期分歧：三年落地还是半个世纪？ Josh 提出，业内绝大多数人（包括他自己）通常认为人形机器人的全面普及还需要 20 年甚至 50 年，而 Andrew 却笃定这将在未来 3 到 5 年内发生。虽然传统工厂早就充斥着各种机械臂，但现在进入视野的是真正具备拟真度、能像人类同事一样协作的人形机器人。Josh 追问，究竟是什么赋予了 Andrew 如此超前的信心？普通人又会在何时、何地真正接触到它们？ Andrew 直言，如果是在 2020 年，他同样会认为这是遥不可及的“登月计划”（moonshot），甚至不确定有生之年能否亲眼见证。但 ChatGPT 问世后，技术的迭代斜率被彻底改写。从 GPT-3 跨越到 GPT-5.5、GPT-6 等代际仅用了三到四年时间，机器智能不仅突飞猛进，在许多领域甚至完成了人类自身无法做到的任务。这种技术演进的指数级路径，将直接复制到“物理 AI”与机器人的智能进化当中。 针对 Josh 提到的大众疑虑——比如网络上疯传的“无头机械狗”令人不安、马斯克带人形机器人去卡戴珊豪宅端酒实则是后台人类远程遥控（遥操）的特技噱头——Andrew 回应称，普通大众要在日常生活中无处不在地接触到人形机器人，确实还需要数年时间；但在硅谷等核心科技聚集区以及互联网前沿动态中，变革的速度远超大众想象，正如 Figure AI 在此前一两个月引发全网病毒式传播的最新现场直播中所展现的惊人突破。 ### [10:36](https://www.youtube.com/watch?v=1Tgw82x46YA&t=636s) 连续工作八天的“Mini ChatGPT时刻”与工业落地 Andrew Kang 谈到了近期机器人领域的一个标志性突破：一台人形机器人执行物流包裹分拣任务，像人类一样翻转包裹并进行分类，这项操作竟然连续无间断地运行了整整八天。他认为这堪称具身智能领域的“Mini ChatGPT 时刻”——大众终于意识到这不是 PPT 或概念，而是真切发生的现实。这段展示既不是人类远程遥控（Teleoperation），也不是从上百次失败的切片中挑出唯一成功的一次剪辑到网上去博眼球，而是真正长时间的自主运作（Autonomous Operation）。 ```mermaid flowchart TD accTitle: 人形机器人连续自主轮班分拣流程 accDescr: 展示人形机器人在物流分拣场景下如何通过交替轮换和自动充电实现连续八天的无缝作业 subgraph Shift["轮班作业流程"] R1["机器人 A：执行分拣包裹与翻转任务"] -->|"工作 8-10 小时后电量低"| C1["返回充电站更换/补满电池"] R2["机器人 B：接替上线继续作业"] -->|"工作 8-10 小时后电量低"| C2["返回充电站更换/补满电池"] C1 -.->|"满电待命"| R1 C2 -.->|"满电待命"| R2 end Shift --> OUT["实现连续 8 天 24 小时自主运行"] ``` Josh Brown 笑着打趣说，自己确实在网上看过那个视频，但也就看了 30 秒，没想到背后竟然持续了一周多；他还开玩笑说还以为机器人中途得请假给老婆打电话或者去空手道班接孩子呢。Andrew 解释道，现场其实是由几台机器人轮流换班：每台机器人工作 8 到 10 个小时，电池耗尽后就自动去充电或更换电池，另一台机器人立即顶替接手。除了分拣包裹，机器人在工厂中还能胜任机床上下料（Machine Tending）、产品包装等极其繁琐的工作。人类能做的事机器人几乎都能做，甚至能胜任那些过于危险、负荷过大或人类身体无法承受的高强度严苛任务。 - **八天自主测试**：无人干涉、无人工遥控，多机轮班换电，证实自主作业稳定性。 - **工厂应用场景**：包裹分拣、物料翻转、机床上下料（Machine Tending）及重载/高危环境作业。 --- ### [12:08](https://www.youtube.com/watch?v=1Tgw82x46YA&t=728s) 惊人的经济账与万亿制造瓶颈：特斯拉的领跑优势 聊到商业化推进，Josh Brown 引用了 Andrew Kang 公开算过的一笔极具冲击力的账：一台机器人的作业产出大约相当于三名人类工人，而机器人的综合使用成本大约仅为每小时 2 美元，相比之下，一名美国工人的综合时薪高达 35 到 40 美元。在这样夸张的成本差距面前，企业采纳机器人的经济驱动力大到根本无法忽视。Josh 提问：既然经济账已经算通了，那么行业普及的唯一障碍是不是只剩下“产能爬坡速度”以及“说服企业高管采购部署机器人”了？ Andrew 指出，在技术演进上，“智能”（Intelligence）层面的突破在未来一到两年内就能彻底解决，如果走进顶尖人形机器人公司的核心实验室，里面展示出的能力已经令人叹为观止、近乎“魔法”（Magical）。真正的核心瓶颈在于物理层面的**制造能力**（Manufacturing）。目前全球头部公司年产能普遍仅在几千到几万台之间；如果要建成年产能达到 100 万甚至 1000 万台的超级工厂，仍需耗费数年时间，预计制造业产能瓶颈将在未来三到四年内逐步突破。谈及竞争格局，Josh 问特斯拉（Tesla）是否在量产规模上占据绝对领先地位，Andrew 毫不犹豫地回答“100%”，并指出特斯拉目前正在规划年产 1000 万台机器人的超级工厂，在制造规模与产业链整合上远甩行业对手。 - **成本对比**：机器人约 $2/小时（相当于 3 名工人产能） vs. 美国蓝领劳动力约 $35 - $40/小时。 - **发展时间线**：认知与具身智能在 1 - 2 年内成熟，规模化量产瓶颈预计在 3 - 4 年内突破。 - **产能对比**：主流人形机器人厂商年产在数千至数万台区间；特斯拉正在规划年产 1000 万台级别的工厂。 --- ### [14:00](https://www.youtube.com/watch?v=1Tgw82x46YA&t=840s) 直面失业危机：为什么这一次“历史经验”会失效？ 对于机器人革命给人类就业带来的深远冲击，Josh 称赞 Andrew 比硅谷很多遮遮掩掩的高管和思想领袖要坦诚得多。很多科技大佬只在私下喝酒时才敢承认机器将彻底替代体力与脑力劳动，而 Andrew 却在公开场合明确指出：AI 和机器人将同时取代物理劳动与认知劳动，全民基本收入（Universal Basic Income, UBI）将不再是乌托邦设想，而是不可避免的社会生存底线。Andrew 直言：“你永远无法抗拒技术潮流，历史上每一个试图对抗科技进步的人都输了。”如果等到两三年后大批人群已经实质性失业才被动应对，全社会都将陷入剧烈阵痛，因此必须未雨绸缪提前建立制度防线。 面对 Josh 提及的技术乐观派论调——“历史上每一次技术革命（比如从铁匠、马夫转变为汽车修理工，或者催生出社交媒体经理）都会创造出更多意想不到的高阶新岗位”，Andrew 提出了极具穿透力的反驳。他指出，过去之所以能诞生新工种，是因为人类把注意力转移到了需要更复杂抽象规划与高级认知思维（Complex Cognitive Thinking）的领域；但这一次截然不同：**AI 和具身智能正在攻克通用思维与认知边界，它具备人类拥有的所有思考维度**。当机器连“思考”和“认知”都能取代时，以往“技术升级催生新就业”的历史外推逻辑将彻底失效。 ```mermaid flowchart TD accTitle: 历次工业革命与AI具身智能革命的岗位替代逻辑对比 accDescr: 对比传统技术革命与当前AI具身智能革命在认知和体力层面的替代差异 subgraph Past["历史技术浪潮（汽车、计算机等）"] P1["替代基础体力/重复劳动"] --> P2["人类迁徙至高阶规划与复杂认知岗位"] P2 --> P3["创造全新工种（如汽修工、社媒经理）"] end subgraph Present["AI 与具身智能革命"] A1["同时替代物理劳动与认知劳动"] --> A2["AI 具备通用思考与复杂规划能力"] A2 --> A3["新工种（如机器人维修）同样可由机器人自我执行"] end ``` 甚至连 Josh 开玩笑提到的“机器人维修工”（Robot Repair Mechanic），Andrew 都冷静地打破幻想：“机器维修为什么不能由机器人自己做？未来绝对是机器人制造机器人、机器人维修机器人。”Josh 随即半开玩笑地哀嚎：“求你别说了，你把大家仅存的几个好工作都给夺走了！” --- ### [18:02](https://www.youtube.com/watch?v=1Tgw82x46YA&t=1082s) 通往后劳动时代：富足社会、UBI 与现实困境 当被问及对普通人的未来是偏向悲观还是乐观时，Andrew 阐明了自己的中立态度：这完全取决于人类能否解决 **AI 安全与 AI 对齐**（AI Safety & Alignment）问题。如果处理得当，埃隆·马斯克（Elon Musk）所描绘的“全民极大富足”（Abundance for Everybody）并非天方夜谭——当劳动力成本被降到极低，整个社会的生产力总和将呈指数级爆发。关键在于人类必须制定出全新的经济价值分配框架，让技术红利能够惠及所有阶层，而不仅仅是资本寡头。 针对全国性政策规划，Andrew 呼吁政府必须从现在开始严肃调研和试点 UBI。虽然欧洲部分地区有过实践且结果喜忧参半，但当社会出现大规模结构性失业、人们根本无工可做时，除了无条件发放基本生活保障金之外，政府将别无选择。谈到 OpenAI 首席执行官山姆·奥特曼（Sam Altman）提出的“通用基础智能”（Universal Basic Intelligence）概念，Andrew 务实地指出了其局限性：普通大众不可能凭空把“智能额度”直接变成餐桌上的面包和遮风避雨的住房，政府终究必须提供真金白银的物资与收入支持。Josh 紧接着提到，北欧国家利用主权财富基金（如挪威石油基金）来反哺全社会的治理模式，或许可以为机器人时代的社会分配提供关键参考。 ### [20:17](https://www.youtube.com/watch?v=1Tgw82x46YA&t=1217s) 机器民主化与政府在具身智能中的角色 Josh Brown 提出了一个类比：一些国家（如挪威的主权财富基金或阿拉斯加的永久基金）通过建立全民繁荣基金来管理国家石油财富，将红利普及给每位公民，从而避免了一两个人垄断全部资源、其余人只能干看着的贫富悬殊局面。他询问 Andrew Kang，面对机器人可能创造的惊人财富，政府是否也应当介入机器人投资，以确保这种生产力红利被全民共享，还是说这种模式带来的问题会多于解决方案。 Andrew Kang 回应称，美国政府实际上已经在朝着这一方向推进。例如，政府已经通过法案直接注资本土半导体公司（如 Intel），并为稀土加工产业链提供关键融资；机器人行业显然就是下一个战略重点。政界领导人频频在公开场合强调机器人对国家安全和竞争力的极端重要性。尽管美国在激励本土机器人产业的步伐上过去略微落后于某些国家，但当前这已成为最高优先事项，国会和参议院目前正在推进多项相关立法支持。 ### [21:56](https://www.youtube.com/watch?v=1Tgw82x46YA&t=1316s) RoboStrategy 的架构：公募版闭式风投基金 Josh 将话题转向 Andrew 创立的核心载体——刚于五月份在纳斯达克上市的投资实体 RoboStrategy（股票代码：BOT）。与传统 ETF 或私有风险投资基金不同，它被设立为一个上市交易的封闭式基金（Closed-End Fund）。Josh 戏称它身上还带着“新基金的出厂气味”，并请 Andrew 详述其设计初衷。 ```mermaid flowchart TD accTitle: RoboStrategy 业务架构图 accDescr: 展现 RoboStrategy 通过二级公开市场募资并投资一级市场头部私有机器人企业的飞轮架构。 subgraph PublicMarket["公开资本市场 (Nasdaq: BOT)"] PUB_INV["大众公开市场投资者"] -->|二级市场买入 / 溢价发行| BOT["RoboStrategy (BOT)"] end subgraph PrivateMarket["一级私有机器人市场"] BOT -->|风险投资与股权收购| ROBO_PRV["头部私有机器人企业<br>(如 Figure AI 等前沿实体)"] end ROBO_PRV -.->|100~1000倍行业增长潜能| BOT ``` Andrew 解释道，RoboStrategy 的本质就是一家**公开上市的风险投资公司（Publicly Listed VC）**，唯一目标是投资领先的机器人公司，且以未上市的私有企业为主。他与团队坚信机器人行业未来将迎来 100 倍乃至 1000 倍的爆发式增长，人形机器人将全面落地并解锁无数工业与服务场景。然而，最激动人心的技术创新全部发生在一级市场；普通公开市场的散户投资者根本无法触及这些非上市企业。RoboStrategy 的使命就是打破这一壁垒，为大众投资者搭建一座直通顶级私有机器人公司的流动性桥梁。 ### [23:17](https://www.youtube.com/watch?v=1Tgw82x46YA&t=1397s) 溢价发行与微策略模式的估值逻辑 Josh 探讨了 RoboStrategy 与 Michael Saylor 掌舵的 MicroStrategy（微策略）之间的相似之处：在股价相对净资产价值（NAV）出现溢价时增发股份，再利用募集的资金买入更多核心资产（微策略买比特币，RoboStrategy 买机器人企业股权），形成自增强飞轮。但两者的核心差异在于定价透明度：微策略持有的比特币每秒都在进行市价计量（Mark-to-Market），而私有机器人公司的估值通常是按季度由估值机构评估，这构成了理解 BOT 价值的核心逻辑。 Andrew 透露，他们正计划将官网的 NAV 数据改为按月更新。他进一步深入剖析了二级市场与一级市场在估值机制上的本质错位： - **流动性与盯市滞后**：公募资产每分每秒都在反映市场情绪与最新进展，而私有企业往往只能锚定在上一轮融资估值，或者因流动性折价而进一步调低账面估值。在两轮融资之间发生的重大研发突破或利空，可能隔着 6 个月、1年甚至更久都无法体现在账面净值上，这也是 BOT 经常以溢价或折价交易的根本原因。 - **市场参与者广度与估值倍数套利**：公开市场的资金体量和买家基数是私有市场无法比拟的。许多成熟公司（如 TransDigm 或 Constellation Software）的核心逻辑就是利用估值倍数差进行无风险套利——公开市场赋予它们 15 到 40 倍的市盈率估值，它们便以 3 到 8 倍的估值在一级市场并购同类企业，资金一入资产负债表便立刻实现数倍的价值增殖。 Josh 补充道，这种“倍数套利”（Multiples Arbitrage）在财富管理行业（RIA）极其普遍：一家大型上市顾问机构收购一家年收入 700 万美元的私营夫妻店（Mom-and-Pop RIA），转手就能将这部分收入按照大机构更高的倍数折算进自身企业价值。Andrew 确认，RoboStrategy 实际上是将这种在私募股权投资（PE）中司空见惯的估值套利工具，历史上首次规模化地搬到了风险投资（VC）和公募股票市场上。 ### [28:03](https://www.youtube.com/watch?v=1Tgw82x46YA&t=1683s) 一级半二级市场转让与真实公允价值的博弈 Josh 接着询问估值技术细节：如果在两轮正式机构融资之间，有初创企业员工通过二级股权转让平台抛售股票，RoboStrategy 是否会将这类非公开股份转让的价格纳入估值模型。 Andrew 指出，目前此类交易在估值模型中所占权重极小，未来可能会适当提高。他强调一级市场的次级转让（Secondary Markets）存在非常特殊的行为偏差： - 它们缺乏充分的价格发现机制，通常盲目锚定上一轮融资定价（例如“半年前估值是这个价，我就按这个价出”）； - 交易不具备公开报备要求，信息极度不透明、甚至充满黑盒与噪音。 Josh 举了一个生动的假想案例予以佐证：假设像 Anthropic 这样的独角兽刚刚按 50 亿美元估值完成一轮融资，三周后突然拿下了五角大楼的巨额大单，其真实价值必然暴增；然而，因为没有公开发行的股票代码进行高频撮合交易，除非促成全新的重大机构融资，否则这部分新增的内在价值在账面上就完全无法被精准追踪。Andrew 认同这一现实，这也正是公开市场工具在折射私有创新成果时所面临的独特挑战。 ### [30:24](https://www.youtube.com/watch?v=1Tgw82x46YA&t=1824s) 传统 NAV 框架与机器人估值飞速迭代的脱节 在谈及一级市场与二级市场的脱节时，Josh Brown 指出，物理 AI 与机器人领域的变化快得令人目不暇接，这种极端的不确定性与常规公开市场投资截然不同。Andrew Kang 对此深有体会，并坦言在管理封闭式基金的资产净值（NAV）框架时充满挑战。为了避免高估资产估值标签（Marks），团队在账面评估上始终保持保守谨慎，但私募股权二级市场本身不仅缺乏透明度，而且流动与定价极度迟缓。相比之下，公开交易市场的投资者反而通过赋予基金显著的溢价交易，更为激进且实时地反映了这些底层资产的真实增值。 Andrew 直言，华尔街现行的 NAV 估值体系根本不是为了估值呈指数级爆发的前沿科技而设计的。在历史上，绝大多数封闭式基金配置的都是私募信贷（Private Credit）或地方政府债券（Municipal Bonds），那些资产的净值一个季度才波动一次。然而在具身智能领域，初创企业的研发进展、人才流入和市场估值在短短几个月内就可能发生数倍的剧变，传统的会计标记周期完全跟不上物理 AI 的进化节奏。 --- ### [31:30](https://www.youtube.com/watch?v=1Tgw82x46YA&t=1890s) 集中持仓策略：寻找机器人领域的 OpenAI 与 Anthropic Josh Brown 提到 RoboStrategy 的持仓极为集中，其三大核心重仓标的——Figure AI、Apptronik 以及 Dyna Robotics——各自占据了基金净资产约 25% 的比例，合计权重高达四分之三。面对 Josh 关于“是否为了精选头部而放弃分散投资、从而承受单一标的暴雷风险”的提问，Andrew 解释称，这种超高集中度在一定程度上源于基金一年前成立时的市场环境。当时赛道内足够成熟、具有真正竞争力的公司屈指可数，但随着顶级人才与资本如潮水般涌入，行业的成熟度每个季度都在质的飞跃，未来的持仓比例会随着标的丰富而动态调整。 ```mermaid flowchart TD accTitle: RoboStrategy 投资选择与幂次法则逻辑 accDescr: 展现 RoboStrategy 放弃平庸标的并下注头部具身智能公司的投资决策链条。 subgraph SelectionStrategy["幂次法则（Power Law）筛选机制"] A["拒绝平庸标的<br/>（估值低但技术落后的跟随者）"] -->|淘汰出局| B["面临巨头资源压制<br/>高概率在激烈竞争中失败"] C["精选行业龙头<br/>（具备顶级人才、技术与领导力）"] -->|巨额资本与研发壁垒| D["锁定具身智能领域的<br/>OpenAI 与 Anthropic"] end subgraph UpsidePotential["百倍回报空间"] D --> E["即便目前已达数十亿美元估值<br/>仍具备潜在 100 倍上行空间"] end ``` 在投资哲学上，RoboStrategy 坚决遵循风险投资中的“幂次法则”（Power Law）。Andrew 强调，他们要寻找的是面向全球最大潜在市场、即使当前估值已达数十亿美元仍具有 100 倍增长空间的终极赢家： - **拒绝“便宜”的平庸公司**：在生成式 AI 浪潮中，如果因为 OpenAI 或 Anthropic 估值太高而去买便宜的二流跟随者，最终几乎都会被头部的技术和算力资源无情碾压；机器人领域也将重演这一剧本。 - **阿尔法收益来自识别胜者**：未来人形机器人行业将会出现大量失败清算的案例，真正的超额收益（Alpha）并不来自于撒胡椒面式的分散，而在于精准鉴别并重仓能够活到最后的超级领跑者。 --- ### [33:42](https://www.youtube.com/watch?v=1Tgw82x46YA&t=2022s) 三大核心持仓剖析：Figure AI、Apptronik 与 Dyna Robotics 针对三大核心持仓的具体逻辑，Andrew Kang 逐一进行了深度拆解： - **Figure AI（人形机器人领域的苹果公司）**：Andrew 将 Figure AI 喻为“人形机器人界的 Apple”。它不仅在底层制造业、具身 AI 模型和硬件架构上处于行业天花板，更在于其打造极致产品体验的能力。走入 Figure 的办公室宛如直接跨入未来——上百台机器人在身边自如穿梭，自主完成家务与物流分拣测试。对于未来售价，Josh 猜测第一代面向公众的商业化机型是否会高达六位数（十万美元以上），Andrew 表示最终定价尚未敲定，但团队正努力让其更加平民化；同时，工商业领域将率先普及机器人租赁（Leasing）商业模式，随后再渗透至家庭消费场景。 - **Apptronik（老牌核心动力源转向整机开发）**：拥有近十年的研发积淀，团队包含业内公认最顶尖的执行器（Actuators）创新元老。执行器如同机器人的关节电机，是决定运动自由度与负载能力的核心命脉。Apptronik 早年曾为业内知名的头部机器人企业研发原型机，如今已转型自主研发整机，并构建了强大的生态联盟：制造端联手电子制造巨头捷普（Jabil），AI 算法端则与谷歌 DeepMind 展开深度战略合作。 - **Dyna Robotics（专注于非双足轮式与固定工位机器人）**：与 Figure AI 软硬件全栈自研路线不同，Dyna 聚焦于“无腿化”的实用形态——采用双臂固定工位或轮式底盘移动。Andrew 强调，双足行走的机械关节不仅极其复杂，而且大幅拉高了制造成本。然而在真实的工厂装配线上，绝大多数工序都是在平整地面上进行的高度重复性劳动，并不需要爬楼梯或越野穿越复杂地形，轮式或台架形态足以完美匹配海量工业场景。 --- ### [36:51](https://www.youtube.com/watch?v=1Tgw82x46YA&t=2211s) PIPE 定增增发逻辑：复刻 MicroStrategy 的每股资产增厚模式 Josh Brown 转向了资本运作层面的敏锐话题：RoboStrategy 在此前两天刚刚宣布完成一笔 3650 万美元的私募股权投资（PIPE）增发，他询问投资者该如何理解这种新股发行带来的潜在摊薄问题。Andrew 披露，在过去短短三周内，公司实际上连续完成了三轮 PIPE 融资，募资总额已逼近 1 亿美元大关。 ```mermaid flowchart LR accTitle: RoboStrategy 增厚型定增机制 accDescr: 阐述 RoboStrategy 如何通过溢价定增实现每股基础资产的持续增加。 subgraph PIPEIssuance["高溢价定向增发"] A["以净资产 3 至 4 倍溢价<br/>（接近二级市场交易价）发行新股"] B["引入不抛售的<br/>长线机构投资者"] end subgraph CapitalDeployment["资本高效配置"] A --> C["募集近 1 亿美元充沛资金"] C --> D["加速扫货重仓顶级机器人标的<br/>（如 Figure AI 股权）"] end subgraph AccretiveResult["增厚股东价值"] D --> E["实现每股内含资产大幅递增<br/>（复制 MicroStrategy 模式）"] end ``` Andrew 详细阐明了增发背后的“增厚型投资”（Accretive Dilution）逻辑： - **机构高溢价入场**：所有的定增份额均被极具长线眼光的主权与大型机构投资者认购，且定价均紧随当期二级市场的高交易价，相当于公开报告 NAV 的 3 到 4 倍溢价发行。机构甘愿支付如此巨额的溢价入局，本身就是对底层资产价值的极端看好。 - **管理层利益高度一致**：管理团队本身是公司的最大股东，绝不会做出损害股权价值的恶意稀释。只有当定增带来的高溢价资本流入加上这笔资金所投项目的潜在百倍回报，能够以压倒性优势覆盖稀释成本时，管理层才会启动发行。 - **对标 MicroStrategy 战略**：如同 MicroStrategy 通过发行股票与可转债将其“每股内含比特币数量”（Bitcoin per share）在 2020 年后放大了 5 到 10 倍一样，RoboStrategy 的终极目标也是借助资本运作大幅提升投资者的“每股内含 Figure AI 份额”，通过持续增厚优质资产来为长期持股者创造几何级数的回报。 ### [39:44](https://www.youtube.com/watch?v=1Tgw82x46YA&t=2384s) 破局传统 ETF：捕捉未上市机器人巨头的投资新路径 主持人 Josh Brown 对本期对话的核心投资逻辑进行了提炼与总结。他指出，对于那些坚信机器人与具身智能（Physical AI）未来必将爆发、且希望通过提早布局从中获利的投资者而言，市场现状正面临一个痛点：虽然市面上已有不少公开交易的传统机器人 ETF，但这类产品往往只能覆盖成熟的传统制造业或大型科技股。而 Andrew Kang 所创立并在纳斯达克挂牌上市的 RoboStrategy（股票代码：BOT）则开辟了一条截然不同的路径——它让普通投资者能够提前触达处于 Pre-IPO 阶段、手握顶尖技术和前沿产品的未上市机器人核心企业，共享它们在真正走向公开市场之前的巨大增长红利。 ```mermaid flowchart TD accTitle: 机器人赛道投资路径对比 accDescr: 对比传统机器人ETF与RoboStrategy投资未上市前沿机器人公司的不同模式 subgraph Traditional["传统投资路径 (公开市场)"] A["公开交易机器人 ETF"] --> B["传统工业自动化 / 成熟硬件公司"] B --> C["缺乏前沿具身智能与早期初创的高倍成长性"] end subgraph Modern["RoboStrategy 创新路径 (Nasdaq: BOT)"] D["公开交易载体 (BOT)"] --> E["优质 Pre-IPO 机器人公司 (如 Figure AI)"] E --> F["捕获百亿美元至万亿美元估值爆发期的资本增值"] end ``` Andrew Kang 对此表示完全赞同，并深入剖析了当今科技初创企业的生命周期变化。他强调，如今高估值独角兽和硬科技初创企业“保持私有化的时间越来越长”（staying private longer）。由于当下私募股权市场上沉淀了极为充沛的海量资金，顶尖科技公司在估值达到数千亿甚至上万亿美元之前，根本没有急于上市的流动性压力；不仅如此，IPO 过程不仅冗长繁琐，上市后还要面对来自公开市场董事会、合规审查以及季度业绩预期的多重桎梏。只要具备留在私募市场的资本底气，绝大多数顶尖团队都会优先选择推迟上市，而这也导致普通散户极难直接参与到这一轮具身智能爆发的核心红利期。 ### [40:47](https://www.youtube.com/watch?v=1Tgw82x46YA&t=2447s) 免责合规与未来展望：把实体机器人带进演播室 在节目的尾声，Josh Brown 以半开玩笑的口吻对观众做出了标准的专业财务合规免责声明：“让我来兜个底，保护好我们俩的屁股（cover both of our asses）。”他郑重提醒所有听众与观众，本节目中所探讨的任何内容均不构成购买或出售任何证券的背书或个性化投资建议，投资者务必自行进行独立研究、运用个人独立判断，并认真权衡相关风险。随后，Josh 询问了想要深入研究 BOT、查阅公司底层持仓及风险披露信息的投资者应当前往何处。 Andrew Kang 给出了官方信息渠道，并相约后续继续跟进机器人行业的前沿发展： - **官方网站与研报查阅**：访问 RoboStrategy 官方主页（RoboStrategy.co）阅读详细持仓与招股风险提示。 - **社交媒体动态**：在社交媒体平台 X（原 Twitter）上关注 `@RoboStrategy` 获取最新资讯与动态追踪。 对话在轻松热烈的氛围中收尾。当 Josh 热情询问未来当更多机器人落地商用时能否邀请 Andrew 再次返场做深度追踪时，Andrew 欣然应允并幽默回应：“下一次，我们直接把实体机器人带进演播室来一起录节目！”Josh 随即兴奋地表示非常期待，并向所有收听与观看本期节目的观众道谢告别。 ## Notable quotes > [02:55](https://www.youtube.com/watch?v=1Tgw82x46YA&t=175s) Just Own the Damn Robots — Downtown Josh Brown > [14:00](https://www.youtube.com/watch?v=1Tgw82x46YA&t=840s) 你永远无法抗拒技术潮流，历史上每一个试图对抗科技进步的人都输了。 — Andrew Kang > [14:00](https://www.youtube.com/watch?v=1Tgw82x46YA&t=840s) 机器维修为什么不能由机器人自己做？未来绝对是机器人制造机器人、机器人维修机器人。 — Andrew Kang > [14:00](https://www.youtube.com/watch?v=1Tgw82x46YA&t=840s) 求你别说了，你把大家仅存的几个好工作都给夺走了！ — Downtown Josh Brown > [33:42](https://www.youtube.com/watch?v=1Tgw82x46YA&t=2022s) 人形机器人界的 Apple — Andrew Kang > [39:44](https://www.youtube.com/watch?v=1Tgw82x46YA&t=2384s) staying private longer — Andrew Kang > [40:47](https://www.youtube.com/watch?v=1Tgw82x46YA&t=2447s) 让我来兜个底，保护好我们俩的屁股（cover both of our asses）。 — Downtown Josh Brown > [40:47](https://www.youtube.com/watch?v=1Tgw82x46YA&t=2447s) 下一次，我们直接把实体机器人带进演播室来一起录节目！ — Andrew Kang ## Mentioned Downtown Josh Brown, Andrew Kang, Michael Batnick, Elon Musk, Sam Altman, Michael Saylor, The Compound, Betterment Advisor Solutions, RoboStrategy, Nasdaq: BOT, Figure AI, Apptronik, Dyna Robotics, Tesla, OpenAI, Anthropic, Google DeepMind, Jabil, TransDigm, Constellation Software, Intel, Goldman Sachs, ChatGPT, GPT-3, GPT-5.5, GPT-6 ## Video description On this episode of Live From The Compound, Downtown Josh Brown is joined by Andrew Kang, CEO and co-founder of RoboStrategy (Nasdaq: BOT), to discuss why humanoid robots may be the biggest investment opportunity of the next decade. They cover the rapid rise of physical AI, Andrew's early bet on Figure AI, the economics driving mass adoption, what widespread automation could mean for workers, and why he structured RoboStrategy as a publicly traded investment vehicle to give everyday investors exposure to the robotics revolution. This episode is sponsored by Betterment Advisor Solutions. Learn more at ⁠⁠https://www.betterment.com/advisors ►00:00 - Ad ►01:23 - Intro ►02:55 - "Just Own the Damn Robots" ►07:36 - Robotics Adoption Timeline ►13:59 - What Comes Next for Workers and Employment ►21:55 - The Structure and Investment Thesis 👕 Check out The Compound shop: https://www.idontshop.com 🎙️Listen to our podcasts: TCAF: https://podcasts.thecompoundnews.com/show/TCAF/ Animal Spirits: https://podcasts.thecompoundnews.com/show/animalspirits/ Ask The Compound: https://podcasts.thecompoundnews.com/show/ATC/ The Compound Newsletter: https://www.thecompoundnews.com/subscribe Linkedin: https://www.linkedin.com/company/the-compound-media/ Instagram: https://instagram.com/thecompoundnews Twitter: https://twitter.com/thecompoundnews Tik Tok: https://www.tiktok.com/@thecompoundnews Investing involves the risk of loss. This podcast is for informational purposes only and should not be or regarded as personalized investment advice or relied upon for investment decisions. Josh Brown and Michael Batnick are employees of Ritholtz Wealth Management and may maintain positions in the securities discussed in this video. All opinions expressed by them are solely their own opinion and do not reflect the opinion of Ritholtz Wealth Management. See our disclosures here: https://ritholtzwealth.com/podcast-youtube-disclosures/ The Compound Media, Incorporated, an affiliate of Ritholtz Wealth Management, receives payment from various entities for advertisements in affiliated podcasts, blogs and emails. Inclusion of such advertisements does not constitute or imply endorsement, sponsorship or recommendation thereof, or any affiliation therewith, by the Content Creator or by Ritholtz Wealth Management or any of its employees. For additional advertisement disclaimers see here https://ritholtzwealth.com/advertising-disclaimers. #thecompound #investing #stockmarket #stocks [Source](https://www.youtube.com/watch?v=1Tgw82x46YA)

[2:55] Downtown Josh Brown: Just Own the Damn Robots [14:00] Andrew Kang: 你永远无法抗拒技术潮流，历史上每一个试图对抗科技进步的人都输了。 [14:00] Andrew Kang: 机器维修为什么不能由机器人自己做？未来绝对是机器人制造机器人、机器人维修机器人。 [14:00] Downtown Josh Brown: 求你别说了，你把大家仅存的几个好工作都给夺走了！ [33:42] Andrew Kang: 人形机器人界的 Apple [39:44] Andrew Kang: staying private longer [40:47] Downtown Josh Brown: 让我来兜个底，保护好我们俩的屁股（cover both of our asses）。 [40:47] Andrew Kang: 下一次，我们直接把实体机器人带进演播室来一起录节目！

[2:55] Andrew Kang: Humanoids are essentially productizing physical labor in the same way that LLMs productize knowledge work... If a robot can do anything that a human can do, and it's working 24/7, 365—it never has to sleep, never gets tired, never takes breaks or complains or quits the job... your market size is essentially the entire physical labor market. [7:36] Andrew Kang: They had a humanoid robot doing logistics tasks, sorting packages, flipping them over the same way a human would... and it went on for eight days straight. I think that was a little bit of a mini ChatGPT moment where people started to realize this is real. [31:45] Andrew Kang: We want to basically find the Anthropics or the OpenAIs of the robotics world... We just want to bet on the best and leading companies in the space because there is so much more upside available. [2:55] Downtown Josh Brown: Just Own the Damn Robots

[0:17] Today's show is brought to you by [0:19] Betterment Advisor Solutions. Every RAA [0:22] knows the tension. You don't want to [0:23] turn people away. You don't want to [0:25] require high minimums. and you want to [0:27] help clients who are just getting [0:29] started because that's where the [0:31] long-term relationships begin. But [0:33] here's the truth. Those simple accounts, [0:35] they take a lot of work. Account [0:37] opening, trading, rebalancing. Before [0:39] long, your staff and back office are [0:41] underwater and trying to stay afloat. [0:43] That's why RAAS are turning to [0:45] Betterment Advisor Solutions. It's the [0:47] platform built for segmenting your book [0:49] and streamlining those smaller and [0:51] simpler accounts. The onboarding [0:53] experience is automated and paperless. [0:56] The portfolio management is streamlined [0:58] and tax efficient. The client experience [1:00] is consistent and exceptional. Explore [1:03] what segmentation can do for your firm [1:05] today. Lower your operational lift, but [1:08] keep your standard of service high. All [1:10] with Betterment Advisor Solutions. Your [1:12] biggest regret will be not doing it [1:14] sooner. Learn more at [1:16] betterment.com/advisors. [1:23] Ladies and gentlemen, welcome to Live [1:25] from the Compound. Humanoid robots are [1:27] no longer science fiction. They're on [1:29] factory floors right now, and the [1:31] numbers suggest the next few years are [1:33] when this goes from early adopters to [1:36] everywhere. Global shipments of humanoid [1:38] and quadriped robots grew 250% last year [1:43] to nearly 53,000 units. Goldman Sachs [1:47] projects more than 250,000 [1:50] humanoid robots shipped by 2030 and a [1:54] total addressable market of 38 billion [1:57] by 2035, a number they have just revised [2:00] up sixfold. One investor saw all of this [2:04] coming earlier than almost anyone else, [2:06] bet his own money on it when the entire [2:09] VC world told him not to, and then [2:11] launched a publicly traded fund on the [2:13] NASDAQ in May to let retail investors in [2:17] on the trade. My guest is Andrew Kang. [2:19] He is the CEO and co-founder of Robo [2:22] Strategy, ticker Boot on NASDAQ, a [2:26] veteran crypto investor who made some of [2:28] the earliest bets in DeFi and then [2:30] walked away from the whole thing to go [2:32] all in on physical AI. Welcome to the [2:35] show, Andrew. So happy to have you. [2:37] >> Hey Josh, thanks for having me. [2:39] >> How's that for a buildup? Pretty good. [2:41] >> Yeah, it's excellent. [2:42] >> All right. All right. Now, I understand [2:44] you also battled Ant-Man in the quantum [2:46] realm, or was that a different Kang? I I [2:48] believe that was a different one. [2:49] There's also a different Kang from Micro [2:52] Strategy itself if you're familiar. [2:54] >> Okay. [2:54] >> All right. [2:55] >> All right. Um in in 2017, let me set [2:58] this up. In 2017, I wrote a blog post. [3:02] Um just own the damn robots and it sort [3:05] of went like mini viral in the financial [3:07] blogosphere, which is a thing that used [3:09] to exist. Um, but I was explaining the [3:12] relatively recent at that time NASDAQ [3:16] explosion in terms of people realizing [3:20] that these companies were on the verge [3:22] of automating the entire economy and [3:24] that investors worried about the future [3:27] should just get longte. And after a [3:30] decade in which the NASDAQ 100 is up [3:32] about 500% since then, um, I'd say it [3:36] was one of my better calls. You've taken [3:38] that a step further. You have personally [3:41] and professionally gone all in on [3:42] robots. Tell us about the origin of your [3:47] decision to want to do that. [3:49] >> Yeah. So, the origin of my decision was [3:52] around around two or three years ago. [3:55] Um, you know, a friend of mine showed me [3:56] this company, Figure AI, and I took a [3:59] look at the video and they were having [4:02] robots walking around in a lab looking [4:05] like humans, just like you see in the [4:06] sci-fi movies. And it was like, well, [4:09] before you would have thought that was [4:11] kind of crazy, but now it was in the [4:13] context of chat GBT having launched, I [4:15] think a little bit less than a year ago. [4:17] And it just kind of made sense that, [4:19] okay, these things can probably actually [4:21] start working now. [4:22] >> Um, it's always a pipe dream. [4:24] >> But, you know, the reason why we didn't [4:26] have it work was because intelligence [4:27] was always the gaining factor. And I [4:29] think it was kind of a little bit more [4:31] clear that intelligence development of [4:33] it was really going to accelerate and [4:35] that you could potentially solve general [4:37] intelligence for robotics in a in a [4:39] really reasonable amount of time. You [4:40] know, not 50 years, not 100 years, but [4:42] something more like three to five to [4:44] seven years. And so that made really [4:46] compelling time. [4:47] >> So figure AI had I guess prototypes of [4:51] robots, but this sort of predates the [4:54] rise of generative AI. it's before chat [4:56] GPT or right around the same time and [4:59] that for you that's the thing that [5:00] unlocks it where you say okay I now [5:04] think this is workable and investable. [5:06] >> Yeah, it was late 2023 you know like [5:09] January 2024 or so and um so it was a [5:12] little bit after CHPD came out. Um [5:15] >> Okay. [5:15] >> Yeah. [5:16] >> Okay. Um you've told the story elsewhere [5:21] you you talked to a lot of other VCs. [5:23] you went to your whole network and you [5:25] were talking about figure and robotics [5:27] and you say almost everybody told you [5:30] not to invest um but then you put in a [5:33] million bucks anyway then you raised [5:35] that stake to 19 million. Walk us [5:38] through that period of time. What do you [5:40] think you saw that the rest of them [5:42] didn't? And are you very happy to have [5:45] not listened uh to to that advice given [5:48] the the rise of robots as an investment [5:50] theme you know within a couple of years [5:52] later? Yeah. Um, so just kind of like [5:56] looking at the company figure and trying [5:58] to understand the total market [5:59] opportunity of it. It it was it was it [6:03] was a new framework that you had to kind [6:04] of have kind of had to develop because [6:06] we never had humanoids before. And so [6:08] there was an established market for this [6:10] that you could look at. um completely [6:13] new and and so like the way we kind of [6:14] thought about it was well humanoids are [6:17] essentially productizing you know [6:19] physical labor in the same way that LLMs [6:21] productize knowledge work you know [6:24] digital intelligence and so you know if [6:27] a robot can do anything that a human can [6:29] do uh and it's working 24/7 365 it never [6:33] has to sleep it never gets tired never [6:35] takes breaks or complains or quits the [6:37] job right like [6:39] >> no health insurance [6:40] >> pretty good [6:40] >> no 401k Okay. Match, right? [6:43] >> Yeah. And like then so then your your [6:46] market size is essentially the entire [6:48] physical labor market which is something [6:50] like $50 trillion. I know that's a [6:52] really nebulous number. It's just kind [6:53] of like a [6:54] >> really big thing. And so it's how much [6:56] perspective 50 trillion [6:58] >> 50 trillion is the annual value of [7:02] physical labor in our economy, [7:05] >> right? It's all the people Yeah. I mean [7:07] that are you know working you know [7:09] factory jobs or uh you know working in [7:11] retail environments but it's also but it [7:13] doesn't include all of the labor gaps [7:15] that we have and all the economic value [7:18] that could be provided by you know all [7:20] the jobs that exist but just can't be [7:22] filled as well. And so [7:25] the market could potentially be larger [7:26] if we start bringing robots to space uh [7:30] having them in other environments where [7:31] it just doesn't make sense economically [7:32] for them to you know have people doing [7:35] those jobs today. [7:37] you are different than a lot of people, [7:39] myself included. I always think like, [7:41] yeah, robots 50 years from now. Um, and [7:44] you know, or a lot of people would say [7:46] 20 years from now, but you're like a 3 [7:48] to 5 years from now person. And I think [7:51] that's where you stand out from other [7:53] pe. I think conceptually we all [7:55] understand there have been robots in [7:57] factories for a long time. The big [7:59] difference now is that they're humanoid [8:01] robots. They almost they they don't just [8:03] act as a coworker. they sort of there's [8:05] a verse similitude with a real actual [8:08] coworker, but you don't think that takes [8:10] 20 or 50 years. You think that's like [8:13] right this minute? Um, [8:16] what what makes you confident that that [8:18] will be the case? And where do you think [8:20] our audience will look around and see [8:22] robots next to them the soonest? Yeah, [8:25] look, I I think in 2020 I I probably [8:28] would have been on the same page and [8:30] thought that this was something that was [8:32] more of a moonshot and, you know, wasn't [8:34] sure if this would be achieved in our [8:36] lifetime. But with when CH GPT came out, [8:38] I I think it was pretty apparent that [8:40] the slope of technological process was [8:42] changing quite dramatically and things [8:44] would be getting better much much [8:46] faster. And so we saw that jump from [8:48] GPT3 to GPT 5.5 5.6 six right in the [8:52] last three to four years and that was a [8:54] huge jump. I mean these right machine [8:57] intelligences are doing things that are [9:00] sometimes things that we can't even do [9:02] ourselves. Uh right the intelligence [9:04] levels have just skyrocketed and so I [9:07] think you can just easily extrapolate to [9:09] that similar path of development [9:11] happening for physical AI as well for [9:13] the intelligence of robots. you have a [9:16] you you have a lot of um you you have a [9:19] lot of contacts in the space and you've [9:21] seen a lot more than the average person [9:24] who's listening or watching us right [9:25] now. I think like speaking as the voice [9:28] of the average person, everyone's been [9:30] sent the YouTube video of the dog with [9:33] no head, which is very disconcerting. [9:36] Um, everybody remembers when Elon Musk [9:39] brought one of his humanoid robots to [9:42] the Kardashian mansion, um, to, I guess, [9:45] serve drinks at a party. And then the [9:47] thing that they say is like, "All right, [9:49] the dog looks like a toy. Um, and the [9:53] Elon Musk stunt with the robot, there [9:56] was actually a person controlling the [9:58] robot behind the scenes." like I still [10:00] feel that people don't really believe [10:02] that this exists or or that they're [10:05] going to run into one of these. So, just [10:07] to like double double down on my [10:08] original question, [10:11] where would you tell uh regular people [10:14] that they're going to come into contact [10:15] with these humanoids first? [10:18] >> Yeah. I mean, in regular contact, I [10:20] would say it's probably a few years out [10:21] from now. Uh, but I think you're going [10:23] to see it more so on the internet or if [10:26] you live in specific places like Silicon [10:27] Valley a lot sooner. Um, I don't know if [10:30] you saw the figure AI live stream uh was [10:33] around one or two months ago that went [10:34] pretty viral where they had a humanoid [10:36] robot doing, you know, logistics task uh [10:40] sorting packages, flipping them over [10:42] like the same way a human would [10:43] >> and it went out went on for eight days [10:45] straight. And I think that was like a [10:47] little bit of a mini chat moment where [10:48] people started to realize this is real. [10:50] It wasn't controlled by a human and [10:52] right it didn't need to it wasn't some [10:54] clip that was chosen out of a hundred [10:56] clips or 100 takes right and that was [10:58] the one that was posted on the internet [10:59] but right it was working autonomously [11:01] for a long time and [11:04] >> wait the robot was sort I've seen that [11:06] clip but I watched 30 seconds of it I [11:08] didn't watch eight days of it but the [11:10] the robot spent over a week at that task [11:13] I guess [11:14] >> yeah they had a few they had a few [11:16] robots changing shifts because each [11:18] robot after you know I don't eight 10 [11:20] hours they had to go and charge their [11:22] right batteries and there was another [11:24] robot that came and swapped over. [11:25] >> I was going to say they had to go call [11:26] their wives or or pick up their kids at [11:28] karate or something but all right so [11:30] they have to swap the batteries but all [11:32] things all things being equal that's [11:34] what we're going to see in that package [11:37] sorting role for example in like a [11:39] factory setting or a warehouse setting. [11:42] >> Yeah. Package sorting, machine tending, [11:45] uh packaging uh items, I mean all all [11:48] sorts of things. It's it's it's kind of [11:51] hard to kind of list off everything that [11:53] a robot could do in these factories [11:54] because it's just, you know, almost [11:56] everything a human can do, but also [11:57] more, right? Because some tasks are just [11:59] too dangerous or, you know, they require [12:02] too much intensity or strain from a [12:04] human. [12:05] >> That's where a robot can do a really [12:07] great job. [12:08] >> You've done the math on this publicly. [12:10] You said a robot doing the work of three [12:12] humans at roughly $2 an hour versus $ 35 [12:15] to $40 for a US worker. um at that math [12:19] the economic incentive for adoption [12:21] becomes almost impossible to ignore. So [12:25] I guess the only real holdup is how fast [12:29] can you manufacture robots and how [12:32] quickly can you can you convince [12:34] business leaders to insert these [12:37] machines in their in their uh work [12:40] environments? Like what what else really [12:43] is there that that would hold this up? [12:45] >> Yeah, I would say intelligent. Sorry. I [12:47] mean, I would say the manufacturing is [12:48] the main bottleneck. The intelligence is [12:50] going to get there in the next one or [12:51] two years. If you kind of like take a [12:53] look inside the research labs or some of [12:55] these uh headquarters of human motor [12:57] companies, it's it's kind of magical in [12:59] terms of like what they what they're [13:01] able to do. Um, but you know, we can [13:03] only produce so many robots per year [13:05] right now. And to produce say a factory [13:08] that can manufacture a million or 10 [13:10] million robots per year, that'll take, [13:12] you know, a little bit of time to get [13:14] going. So that'll be the bottle, but [13:17] we're going to solve it in the next, I [13:18] would say, three, four years. [13:19] >> So we're going to get into some of the [13:21] investments that you're making today. [13:22] Um, but just conceptually, [13:26] is it likely that Tesla will be at least [13:30] in the early stages of this, the first [13:33] company who actually has the wherewithal [13:35] to manufacture humanoid robots at some [13:38] sort of scale? Do they look like they're [13:40] in the pole position from a [13:42] manufacturing standpoint? 100%. I think [13:45] they're working on a 10 million robot [13:47] per year facility right now, which would [13:49] be the [13:49] >> the biggest one by far. Uh most [13:52] companies are producing on the scale of [13:54] thousands to tens of thousands of robots [13:56] per year today. [13:58] >> So, it's a gigantic. [14:00] >> Um Andrew, you've been very honest about [14:03] what you think comes next for workers, [14:06] maybe more so than a lot of other people [14:08] building in this space and a lot of [14:10] thought leaders in Silicon Valley. Um, [14:13] you think that or you've said AI and [14:17] robotics will eventually replace both [14:20] physical and cognitive labor and that [14:23] perhaps universal basic income or UBI [14:26] becomes a necessity. [14:29] Most executives um in the space again [14:32] they're not saying that out loud. I [14:34] would imagine maybe they're saying it [14:35] over drinks with uh with their [14:37] colleagues, but um you seem to lead with [14:40] it. So talk about why that component of [14:43] it is really important for us to be [14:45] honest about. [14:46] >> Yeah. Well, I mean, you can't fight [14:48] technology. Anyone that's tried to fight [14:49] it in the past, they've always lost. Um, [14:52] and and and so I think you got to start [14:54] planning ahead of time for when that [14:55] reality hits and people are displaced by [14:58] by this technology. Um, and you know, we [15:02] can't do it two years, three years in in [15:04] the future because people might be out [15:05] of jobs at that point. So we need to [15:07] start working on a plan now um for when [15:10] that moment does come because when it h [15:11] comes it's going to happen real fast and [15:13] you don't want to react late and you [15:14] know people could be a lot in a lot of [15:16] trouble if you are reacting late. [15:18] >> The techno optimists seem to think that [15:20] there's going to be a boom in jobs we [15:23] never could have conceived of before [15:25] >> happening in the wake of this robotics [15:28] revolution. And I don't know if that [15:30] means robot repair. I'm not a thousand% [15:33] sure what what that means, but I I guess [15:36] there are a lot of people working in [15:38] jobs today that we couldn't have [15:39] imagined 20 years ago, [15:41] >> like social media manager. So, um, talk [15:44] a little bit about I guess why you're [15:46] hopeful that we're not talking about UBI [15:49] being the way everyone gets paid, maybe [15:52] just a subset of workers who are the the [15:55] first among the displaced while they try [15:57] to figure out what they're going to do [15:58] next. Or is it more bleak than that? [16:01] Yeah, I mean I'm sure there's going to [16:02] be some jobs that uh people will move [16:04] into that people don't want robots doing [16:06] um service related jobs where you want a [16:08] human touch for example and uh you know [16:11] maybe that gives a higher level of [16:12] service if you have a human doing that [16:13] for you as opposed to a robot for [16:15] example in China uh they have the food [16:18] delivery robots in a lot of the hotels [16:21] but they don't have them in the fivestar [16:23] hotels the best hotels right because it [16:24] kind of adds a little bit of like a less [16:26] premium service [16:28] >> but at the same time like what if you [16:30] get androids, right? What if you have [16:31] robots that are almost indistinguishable [16:33] from humans and provide even a better [16:35] level of service? Then well, that's [16:37] maybe pretty tough to to deal with. I [16:40] think the argument goes like this for [16:41] the maybe the some of the technology [16:43] people, which is we've always found like [16:45] new jobs for people, but the reason why [16:47] we found new jobs is because we found [16:49] jobs to do that required more planning [16:51] or like con like complex cognitive [16:54] thinking in some way. And like [16:57] intelligence is AI is not going to stop. [16:59] it's going to be able to think in all [17:01] the ways that humans can. And so I don't [17:04] see that being right like a good [17:06] extrapolation to the future. Um, and [17:10] yeah, I'll just I'll just leave it at [17:12] that. like I [17:14] >> robots. You think so you think these [17:15] previous technological waves where [17:18] within a couple of years somebody that [17:20] used to be a blacksmith or somebody that [17:22] used to um manage a horse farm then we [17:25] get the automobile those people maybe [17:28] become automobile mechanics like that [17:30] that analogy [17:32] may not hold true this time because of [17:36] the thinking that AI is capable of doing [17:40] and that's the thing that prior waves of [17:42] technological advance just it they [17:45] didn't have as a as a feature as a [17:47] feature. [17:47] >> That's a great way to look at it cuz you [17:49] mentioned robot repair mechanic. Why [17:50] couldn't a robot do that? You're going [17:52] to have robots building robots. [17:54] >> Stop. You're take you're taking all the [17:55] good robot repair jobs. Stop. [17:58] All right. I mean, look, it's a serious [18:00] subject. We're laughing about it. Um, [18:02] but it would you say like on the uh [18:05] spectrum of [18:07] uh optimism versus pessimism about what [18:10] this is going to mean for the average [18:12] person, would you say you're leaning [18:14] more toward pessimism or are you just [18:16] neutral? Um because it doesn't sound [18:18] like you're in the 8 n 10 level of the [18:22] optimists who say this is going to be [18:24] incredible because think of all the new [18:25] industries we're going to create. It [18:27] doesn't sound like you're you're you're [18:29] on that side of the the spectrum of of [18:32] thought on this. [18:33] >> Yeah. Well, I think if we get like [18:34] figure out the AI safety AI alignment [18:36] piece, then there could be some pretty [18:38] good outcomes for humanity. uh you know [18:40] Elon talks about embodans for everybody [18:43] because you know labor is just going to [18:45] be so cheap and uh we're going to be [18:46] able to do so much as society and if we [18:49] write the right frameworks for letting [18:52] that benefit all the economic value [18:54] benefit people of you know all classes [18:56] then I think we could have a great [18:59] outcome. Um but we got to get that [19:01] right. Right. [19:02] >> I I agree. What would you do if you had [19:04] the power to enact some sort of plan or [19:08] program at a national level? What would [19:10] you do today or what would you start [19:12] talking about today so that people are [19:16] asking the right questions and looking [19:17] for the right answers to those [19:19] questions? [19:20] Well, I think there needs to be more [19:22] research done and studies and I think [19:24] there's been some already around UBI and [19:26] what happens to society once you start, [19:29] you know, giving out essentially like [19:31] government support paychecks to to [19:33] everyone. Um, you know, there are some [19:35] societies that have done it in in Europe [19:38] and it's worked out well for them, but [19:39] you know, there have been some [19:40] implementations that haven't worked out [19:42] so well. But it's it's also you don't [19:44] really have a choice. Um because if [19:46] there are no jobs for people then what [19:48] what else are you going to be able to [19:49] do? Um you know Sam Alman I think talks [19:52] about universal basic intelligence. Um [19:55] but I'm not sure like that's a little [19:58] bit harder to translate because I'm not [19:59] sure anyone [20:00] >> will just be able to turn intelligence [20:02] into food and you know housing for [20:05] themselves as well. [20:06] >> Um so you probably got to provide that [20:08] for a lot of people. So one of the [20:10] models that I I guess would be deemed [20:13] successful is in the Nordic countries [20:15] where they've taken the oil wealth of [20:17] the nation and they've created sort of [20:19] this prosperity fund that manages that [20:23] oil wealth and every citizen effectively [20:27] gets a dividend from that. And you know, [20:30] it's it's sort of how they've been able [20:31] to make it so they don't have one or two [20:34] really rich people um who who have [20:36] captured all of that oil wealth and [20:39] everybody else in the country just looks [20:40] at it. It's like, "Oh, that must be nice [20:42] for them." So, it's sort of it's sort of [20:44] an example. I mean, it goes into like [20:46] socialism and areas that you and I [20:48] aren't going to go into today. Um, but [20:50] do you think do you think that might be [20:53] some sort of a model where the [20:55] government is involved in robot [20:58] investments and is able to capture some [21:02] of that uh shared prosperity and make [21:05] sure that it's being sort of universally [21:08] um applied or or is that more [21:11] problematic than it is a solution? [21:14] >> It seems like the government's kind of [21:15] going that direction already, right? [21:16] They've um you know invested in [21:18] semiconductor companies in the US like [21:20] Intel. They provided financing for the [21:22] rare earths uh processing companies and [21:25] I think it's pretty clear that robotics [21:27] is next. They've you know let's been on [21:29] TV over and over saying about how [21:31] important robotics is for the country. [21:33] Um, and you know, we've been a little [21:34] bit behind on it compared to other [21:36] countries in terms of incentivizing the [21:37] domestic industry, but it now it is a [21:40] major priority for the government and [21:42] you're seeing that with the [21:43] conversations that are going on with [21:45] some of the bills that are going through [21:46] Congress and the Senate right now. And [21:48] so, yeah, I absolutely think that um, [21:50] you know, having government investment [21:52] in the industry is is is important. [21:56] >> I want to talk about the structure of [21:57] what you built because this is where it [21:59] gets really interesting for our [22:00] audience. So, you've got a vehicle, [22:03] which you'll describe to us, that went [22:04] public on NASDAQ uh in May. So, it still [22:08] has that new fund smell. Um, the ticker [22:11] is BOT, bot, and it's a closedend fund [22:15] as opposed to an ETF or a venture [22:18] capital fund. It's it's something a [22:20] little bit different. Can you explain [22:22] what Robo Strategy actually is, why you [22:25] structured it this way? Yeah, Robo [22:27] Strategy is essentially a publicly [22:29] listed venture capital company. It is [22:31] exclusively focused on investing in [22:33] leading robotics companies and [22:35] predominantly private robotics [22:36] companies. Um, and really the concept is [22:40] we're really bullish on the industry. We [22:42] think it's going to grow 100fold, a [22:44] thousandfold. Um, we're going to bring [22:47] humanoids, right, to the market. We're [22:49] going to invent all these new [22:51] applications that can be done by robots. [22:53] But all that exciting innovation is [22:55] happening in the private sector. Um, and [22:58] you saw this with the AI lab, [23:00] >> the private markets, the private away [23:02] from away from this. A lot of it's not [23:04] on the the stock market. So, you're [23:07] giving people away to own stakes in the [23:10] private companies that you identify as [23:12] being worth investing in. [23:14] >> That's right. Yep. Exactly. [23:17] >> Okay. Um, the micro strategy comparison [23:21] is something that you've leaned into [23:22] publicly. The idea is that, and if I [23:25] have any of this wrong, just correct me. [23:27] You issue shares at a premium to NAV. [23:31] You use the proceeds to buy more equity [23:35] in robotics companies. And it sort of [23:39] sounds like the flywheel that uh Michael [23:42] Sailor first proposed when he began to [23:45] use Micro Strategy as a way to do that [23:47] with Bitcoin. Um Strategy's NAV is a [23:51] live Bitcoin price effectively like how [23:54] I think he owns uh was it 850,000 [23:57] Bitcoin or or something like that. And [24:00] so the the price of the security is a [24:02] function of where is the price of [24:04] bitcoin and how how many bitcoins does [24:07] strategy own. So it's it's sort of easy [24:09] to understand by people and then it'll [24:11] trade at a premium or a discount to to [24:14] that amount of bitcoins. Um, [24:18] your private company valuations are [24:21] estimated quarterly and that's what [24:23] gives investors a fundamental view into [24:26] what bot could or should be worth at any [24:29] given time. Do I have that right? [24:31] >> Yeah. So, we're we're actually going to [24:34] probably start updating it monthly on [24:36] our website. Um, I think there's a lot [24:38] of nuance in understanding NAV because [24:41] uh, it's very different from the NAV of [24:43] a fund that holds publicly traded assets [24:46] where those assets are pretty much [24:47] marked to market every second or every [24:49] minute for something like Bitcoin. For [24:51] us, right, these private assets are [24:53] marked at the last round valuation or [24:56] sometimes at at a discount. And and so [24:59] what that doesn't take into account is [25:01] all the changes in the company, right, [25:03] between rounds. And like a public [25:06] company that valuation or the growth of [25:09] the company uh is being reflected pretty [25:12] much every single day [25:13] >> every se every second right [25:14] >> that's not reflected in in the NAV and [25:17] so right but if there's a positive [25:18] development or a negative development [25:20] it's not reflected until the next round [25:22] and that could be six months out a year [25:24] out two years out uh and that's part of [25:27] the reason why you might see differences [25:29] from the company right strategy trading [25:31] at a premium or at a discount and maybe [25:34] to kind of reflect reflect some of some [25:35] of that development that might occur [25:37] between rounds. Um, and there's so and [25:39] and and I think you you should also [25:41] understand that private markets and [25:43] public markets value assets differently [25:46] because there's there are two mark there [25:49] are two different markets and they have [25:50] a different subset of market [25:52] participants. And you see this happen [25:54] all the time. More private companies [25:57] when they go public, some of them might [26:00] be valued more highly when they go [26:01] public because there are more buyers [26:04] there. There's more demand in in the [26:06] public markets than there are the [26:07] private markets, right? Which is really [26:09] like a small subset of the total [26:11] investors in the entire world where the [26:13] public markets are a subset of almost [26:15] every investor in in in the world. And [26:18] um and you see this in in some public [26:20] companies where their whole strategy is [26:22] based on rolling up private companies [26:25] into their balance sheet. Companies like [26:26] Transdime or Constellation Software [26:29] where the public companies um you know [26:31] the public markets mark them at [26:33] something like a 15 to 40x earnings [26:35] multiple and their whole strategy is [26:37] buying companies in the private markets [26:39] that are valued at maybe 3 to 8x that do [26:41] very similar things. And so they can [26:44] spend 100 million and as soon as that's [26:46] on their balance sheet now it's 400 [26:48] million right they've created 300 [26:49] million [26:50] >> it's an arbitrage they they do that in [26:52] my industry the larger the larger [26:55] registered investment advisory firms [26:57] have sort of like an enterprise multiple [27:00] >> and so when they go and buy a mom and [27:01] pop RAIA that's immediately like it [27:05] immediately adds to the value of the [27:07] enterprise at like 3 to 4x just by [27:10] virtue of this is now no longer longer a [27:13] standalone business doing 7 million in [27:15] revenue. We're adding this 7 million in [27:17] revenue to us and therefore the multiple [27:20] on that revenue is higher because our [27:22] multiple is higher. This is a very [27:23] common thing. So you're doing that in a [27:26] in a market where um there's less price [27:29] discovery. There are less investors [27:31] total. It's sort of like um in an [27:34] enclave, you know, primarily in Silicon [27:37] Valley or in Asia. And as a result, [27:40] there's some sort of I don't want to [27:42] call it alchemy, but there's some sort [27:43] of arbitrage where once these companies [27:46] become part of your fund, all of a [27:49] sudden there could be an uptick in what [27:51] they're worth. [27:52] >> Yeah. And in private equity world, [27:54] right, they call it uh multiples [27:55] arbitrage. There's a bunch of terms for [27:57] it, and we're just applying to venture [27:59] capital, which is, I think, the first [28:00] time people have done it in in in the [28:02] public equity markets. [28:03] >> Okay. Um, could you update the value of [28:07] the company, the private companies [28:08] you're invested in based on shares [28:11] changing hands amongst private [28:13] shareholders? For example, you invest in [28:16] a robotics company and an employee [28:19] manages to sell some of his or her stock [28:22] to a venture capital fund or through one [28:25] of these uh platforms and then that the [28:28] value is different from the last round. [28:32] Are you able as the manager of the fund [28:34] to take that into account in how you [28:36] value these startups or is that an area [28:38] where you don't want to go you'll just [28:40] wait for the next financing round to to [28:43] write the value up or down? [28:45] >> Yeah, right now it's a small component [28:47] of our valuation framework. [28:49] >> Um could be a larger component over [28:51] time. Uh what I would also note is that [28:54] the secondary markets for venture are [28:56] are pretty funny and that they don't act [28:58] like the public secondary markets. It [29:00] also anchor to, you know, the most [29:02] recent rounds because people say, "Hey, [29:03] look, they just did this round six [29:05] months ago. I want money that way." [29:07] Yeah. And then it jumps, right? Like [29:09] whenever there's a new round, that [29:10] happens. That's a little bit of alpha [29:12] actually, you know. Um [29:14] >> I don't I don't think that the the the [29:16] markets are entirely uh correctly priced [29:20] a lot a lot of the times uh in the [29:21] secondary markets. [29:22] >> Well, it's hard like take um take for [29:24] example Anthropic. Uh so let's say they [29:28] do a round and you know they're raising [29:31] at a valuation of this is like years [29:34] ago. They raise at a $50 billion [29:36] valuation and then news comes out like 3 [29:39] weeks later that they got this massive [29:41] contract with the Pentagon. [29:43] >> Obviously the company's worth more [29:46] >> but there's no ticker and no one's [29:48] trading it. So the private market knows [29:51] it's worth more, but until there's a new [29:53] transaction, hard to say hard to say [29:56] exactly how much more. And that's a [29:58] really big difference between public [29:59] market investments and private. [30:01] >> There's a lot of transactions in the [30:02] secondary market, but the thing is that [30:05] there's no like they're not publicly [30:06] reported and and so there's a lot of [30:09] nontransparency opakquakeness in terms [30:11] of where these things should actually be [30:13] priced and what people are buying and [30:14] selling them at. [30:15] >> Okay. Okay. So, you're being very honest [30:17] and open with your investors in the bot [30:21] fund. Hey, the these valuations are [30:24] going to change rapidly. We hope they'll [30:27] change positively, but you know, they'll [30:30] it'll be sudden and we won't know for [30:32] sure. And I I feel like you're I read [30:35] through some of your stuff. You're [30:37] you're just saying like this is very [30:38] different than public market investments [30:40] and you should, you know, act [30:42] accordingly. [30:43] >> Yeah. It's it's interesting. uh kind of [30:46] navigating this NAV framework because [30:48] it's you know of course we try to be [30:50] conservative on it because you know we [30:52] don't want to be overestimating our our [30:54] our marks um and the secondary markets [30:57] can be intransparent and maybe a little [30:59] bit slow uh at the same time right uh [31:03] like we want to be able to provide [31:04] accurate values and I guess the public [31:07] market is kind of doing that in a in a [31:08] way by trading these things at at a at a [31:11] premium um I hope it's something that [31:14] becomes more transparent over time. But [31:16] the the NAV framework was just never [31:17] made for this world where the valuations [31:19] move so fast, right? It was most of the [31:21] closing funds in the past have been [31:23] private credit funds where right the [31:24] changed quarter [31:25] >> or or they'll own municipal bonds or [31:27] they'll own Right. Okay. Um your three [31:30] largest positions are Figure AI, [31:32] Aptronic, and Dino Robotics. Um [31:36] according to what I read, each [31:37] represents a quarter of the fund's net [31:39] assets. It's a pretty concentrated bet. [31:43] Um it seems like you want to own a lot [31:45] of the best and not just sprinkle in a [31:48] hundred different startups for the sake [31:50] of saying you're diversified. Um of [31:53] course what comes along with that is [31:55] increased risk if something goes wrong [31:57] with one of them. Tell us about those [31:59] companies and um what people need to [32:02] know just like uh [32:05] as far as like why you chose those [32:08] specific investments. [32:09] >> Yeah. So, so one thing I would note is [32:11] that the percentage allocations are [32:13] going to change pretty significantly [32:15] over time. Uh the reason why, you know, [32:17] they might be a little bit more [32:18] concentrated was because when we started [32:19] the fund around a year ago, right, there [32:22] wasn't as many exciting companies in the [32:24] space that were, you know, more mature [32:26] than there are today. And that's changed [32:29] quite a bit because the field is moving [32:30] so fast. There's so much more talent [32:31] going into the space and companies are [32:34] maturing, right? Um, you know, quarter [32:36] by quarter. And uh you know in terms of [32:38] the companies that we've chosen you know [32:40] we focus on companies that have just [32:41] have the biggest power law potential [32:43] impact uh companies that have uh you [32:46] know tackling the biggest markets in the [32:48] world and still even at multi-billion [32:50] dollar valuations have potential 100x [32:53] upside uh right we want to bas basically [32:55] find the anthropics or the open AIs of [32:58] the robotics world and we also want to [33:00] be selective because imagine if you [33:02] invested in maybe a lower valuation like [33:06] AI I company that you felt because you [33:08] felt anthropic or OpenAI was too [33:10] expensive and you were trying to catch [33:12] up to them. I mean, you would have [33:14] probably not done too well because most [33:16] of them, right, they just kind of got [33:17] out competed by these larger companies [33:20] that had more talent, more resources, [33:21] and just were better better leadership [33:24] and and so that's kind of our approach [33:25] where we just want to bet on the best [33:28] and leading companies in the space [33:29] because there is so much more upside [33:31] available. you don't need to be picky [33:32] and uh there's a lot of alpha in being [33:35] able to to differentiate between you [33:37] know the winners and and the losers [33:39] because there are going to be a lot of [33:40] failures in the robotic space. [33:42] >> Okay. So what's the elevator pitch on on [33:44] figure AI? [33:46] >> I mean they're I would think about them [33:48] as like the apple for humanoids. Um they [33:50] have a [33:51] >> they have the best product. [33:53] >> Yeah. It's not just, you know, their [33:54] chops in manufacturing AI and hardware, [33:57] which I believe they're all top of the [33:59] line for, but it's also the experience [34:01] that they've been able to show with the [34:03] robot with, you know, the videos that [34:04] they've created. And, uh, you know, you [34:06] walk into the figure office, it feels [34:08] like you're walking into the future. Uh, [34:10] there's, you know, hundreds of robots [34:11] walking around doing household tasks or [34:13] doing tasks in like a logistics [34:15] environment. And, you know, it's like [34:17] designed in a way where it's like this [34:18] is a premium product. If I have people [34:20] in my home, that's like that's something [34:22] I would show off. I want to I wouldn't [34:24] want to show off like a dinky, you know, [34:26] $5,000 robot that is, you know, like [34:28] breaking, you know, every six months or [34:29] >> what are they going to sell the first [34:31] the the first model that's for sale to [34:32] the public? What are they going to sell [34:34] that for? [34:34] >> Uh I think that's TBD. [34:36] >> Okay. [34:37] >> No, [34:37] >> but we're thinking it's six figures. [34:39] >> Uh [34:41] >> or not necessarily. [34:42] >> It they might have a model that goes [34:44] that high. I think they're probably [34:45] going to try to make it more affordable. [34:47] uh and I think the leasing uh the [34:50] leasing uh robot type of uh business [34:52] model is going to be popular as well [34:54] that a lot of these companies are [34:55] exploring but that's more so to to [34:57] industry rightosed [35:00] for the home. [35:01] >> What's Apptronic? [35:02] >> Appronics uh they were one of the [35:04] earliest humanoid companies. They've [35:06] been around uh for you know nine ten [35:08] years and they have uh what I consider [35:11] some of the original innovators in [35:13] actuator development. Actuators are [35:15] basically the motors in the in the [35:16] joints of the humanoid that basically [35:18] allow them to move around. Uh and you [35:20] know that's one of the most important [35:22] components of of a humanoid or any robot [35:25] in general and they've partner [35:26] >> so that's a bet on a supplier to the [35:28] robot companies. [35:29] >> They're so they they make their own [35:30] humanoids now. Before they were [35:32] consulting service and they had actually [35:33] produced the prototypes from for one of [35:35] the very well-known humanoid companies [35:37] today. Uh but then they decided hey look [35:40] we're going to develop our own humanoids [35:41] now. And so they developing their own [35:44] hardware. They partnered with Jable to [35:45] do the manufacturing. They partnered [35:47] with Google deep mind to develop the AI. [35:49] >> Okay. Uh and lastly, Dino Robotics. [35:52] What's the bet there? [35:54] >> So Dina is a really unique case where a [35:57] company like Figure, they're developing [35:58] both the AI and they're also developing [36:00] the hardware. Um unlike I guess [36:03] electronic is that they're doing it kind [36:05] of entirely vertically within the [36:06] company itself. But their robot doesn't [36:09] have legs. you know, they're creating [36:11] robots that are completely stationary. [36:13] They're just two arms or they're on a [36:15] wheelbase. And I think that's a really [36:17] big market as well because not every [36:19] single robot needs to move around to [36:20] have legs and those add a lot of cost to [36:22] the robot, right? Most tasks in a [36:24] factory, you're just kind of stationary. [36:26] You're doing the same job over and over [36:27] and over again. It's repetitive and [36:30] right you you for a lot of environments [36:32] as well, you don't need to go up and [36:33] down stairs or, you know, go across [36:36] tough terrain, right? You just need to [36:37] be able to move around the same floor. [36:39] >> Yeah. depending on what it depending on [36:40] what its job is. Not every job requires [36:43] like [36:45] mobility on foot. [36:47] >> Yeah, exactly. That's what they're [36:50] working on. [36:51] >> Okay, last question. Um, you just [36:53] announced a $36.5 million private share [36:56] issuance two days ago. Is this I guess [37:01] it's not I want to ask you if it's [37:04] dilutive but I'm not sure if that's the [37:05] right terminology because it's a [37:06] closedend fund and ostensibly you said [37:10] you're deploying it like you're going to [37:12] be making a uh an investment with that [37:15] money so it's additive even to the [37:17] existing shareholders of the fund but [37:19] how I guess how should a shareholder in [37:21] bot think about this uh share issuance [37:26] and future share issuances as you find [37:29] new opportunities that you want to [37:30] invest in, you're going to need capital [37:32] uh to do it. I think everyone [37:33] understands that. [37:34] >> Yeah. So, we've announced three pipe [37:38] fundraises over the course of the past [37:40] three weeks uh totaling close to around [37:43] 100 million. [37:44] >> Those were all done with institutional [37:47] investors that we believe are very [37:48] long-term oriented, not selling them [37:50] into the public market. Um, and they [37:53] were all done uh at multiples to the [37:57] publicly stated NAV. Uh, I believe [37:59] around 3 to 4x. Um, and [38:02] >> so these people believe in what you're [38:04] So these people believe in what you're [38:05] building so much that they're paying a [38:08] very high premium over the actual NAV [38:10] that you're now reporting. So that's a [38:12] sign of a lot of confidence. [38:14] >> It was it was uh close to the market [38:16] price of each day that you know we had [38:19] the pipes. [38:20] >> Yes. [38:20] >> Right. Okay. Of course, like we're not, [38:23] you know, as the largest shareholders in [38:26] the company, the management team is not [38:27] going to want to do dilution or issue [38:29] shares if we don't believe it's [38:30] massively accreative, right? The value [38:32] that we're getting from doing the [38:34] issuance itself at a at a multiple plus [38:38] the value of the investments that we're [38:40] making and what we hope to generate from [38:42] the returns has to outweigh the cost of [38:45] dilution substantially for us to [38:47] decision. [38:48] >> Okay. So if you're a shareholder in bot [38:50] and you see an issuance, your attitude [38:54] should not be, oh, I guess this dilutes [38:56] the amount of exposure I thought I had [38:58] to figure AI. The way that you should [39:01] actually think about it is, okay, they [39:03] must have identified some opportunities [39:05] and they want to go after it and that'll [39:08] be more investment in in this portfolio [39:11] that I'm invested in. [39:13] >> Like that would be the that would be the [39:14] way you want people. We're increasing [39:17] the exposure to the companies that [39:18] people see and want in the portfolio in [39:21] the same way Micro Strategy did that for [39:23] Bitcoin, right? They increased their [39:25] Bitcoin per share. That's one of the key [39:26] metrics they looked look at something [39:28] between five to 10x from when they first [39:30] started accumulating it in in 2020. And [39:33] so you know hypothetically if we had one [39:36] figure share per robo strategy share we [39:39] hope to increase that by multiples in [39:42] future as much as we can so people can [39:44] actually right have more exposure on a [39:46] person. [39:46] >> Okay that that's awesome. So, um, [39:49] somebody that's thinking about investing [39:50] in robot in, uh, robotics and they [39:53] believe as I do and as you do that this [39:56] is happening and, uh, you know, it's [39:59] it's it's something that you can [40:01] actually benefit from if you place your [40:03] bets right. There are lots of publicly [40:05] traded robotics ETFs and now what you [40:08] have listed is something completely [40:11] different because you're able to invest [40:13] in pre-public robotics companies that [40:16] have great products and ultimately at [40:18] some point hopefully will themselves go [40:20] public. [40:22] >> That's correct. Yeah. And you know these [40:24] private companies stay private longer [40:25] and longer. SpaceX IPOed at two trillion [40:28] anthropic open AI not going public until [40:30] at least a trillion. And that trend is [40:32] going to keep going for these private [40:33] companies because there's just so much [40:34] capital going in into these markets. And [40:37] it's it's a long process to go public. [40:39] It was it's not difficult. You have to [40:40] deal with public board and all these [40:42] other things. And if they have the [40:44] ability to stay private, most companies [40:46] are going to choose to do so. [40:47] >> All right, Andrew, I really appreciate [40:49] your time today from a uh let me just [40:51] cover both of our asses. Nothing that [40:53] you hear on this show is ever an [40:55] endorsement or a recommendation to buy [40:57] or sell any securities. Of course, uh, [41:00] do your own research, use your own [41:01] judgment. Um, if you want to read more [41:04] about bot and read the risk factors and [41:08] read about the company holdings, Andrew, [41:10] what would be the best place for people [41:11] to go? [41:12] >> Yeah, roboergr.co or on X. Robo [41:15] Strategy. [41:16] >> Okay. RoboStrategy.co or follow Robo [41:19] Strate Robo Strategy on X. Andrew, [41:22] thanks so much for coming by. Can we [41:24] come to you in the future as we see more [41:26] robots and want to talk more? Yeah, we [41:28] we'll bring in robots in the studio next [41:30] time, Josh. Thanks for having me. [41:31] >> I love I love it. Nice to see you. [41:32] Thanks so much for your time. Thank you [41:34] guys for watching and listening. [41:35] >> Take care.