This episode of Deep Dive provides a comprehensive recap of CES 2026, highlighting a pivotal shift in the technology landscape from Generative AI (chatbots) to "Agentic AI" and "Physical AI." The discussion focuses on how AI is moving beyond the screen and cloud into the physical world to become the foundational infrastructure of consumer technology.

The hosts explain that this new era is defined by "agency"—software that proactively anticipates user needs and executes tasks across devices without waiting for prompts—marking a departure from the passive tools of the past few years.

The episode details specific innovations driving this shift, particularly the rise of local AI processing on PCs and smartphones to ensure privacy and speed (the "AI PC" era). It covers major developments in robotics, such as Hyundai's industrial integration of Boston Dynamics and new home assistants from Samsung and LG that aim to be helpful rather than just novelties.

Furthermore, the discussion touches on the explosion of health tech following new FDA guidance, featuring non-invasive monitoring and smart wellness devices, ultimately painting a picture of a future where technology becomes an invisible, proactive, and helpful layer in daily life.

這集 podcast 針對剛落幕的 CES 2026 進行了深入的趨勢分析，重點在於人工智慧（AI）的重大轉型：從單純的「生成式 AI」（如對話機器人）進化為具備主動執行能力的「代理 AI」（Agentic AI）與「實體 AI」（Physical AI）。節目指出，AI 已不再只是雲端上的軟體，而是成為各種硬體設備的基礎架構（Infrastructure）。透過 Lenovo 的情境感知筆電、Samsung 與 LG 的智慧家電生態系等案例，主持人解釋了 AI 如何從「等待指令」轉變為「預測需求並主動執行」，真正實現跨裝置的無縫協作。

除了軟體代理，節目也探討了 AI 進入實體世界的硬體突破。重點涵蓋了 NVIDIA 與現代汽車（Boston Dynamics）在工業與居家機器人上的進展，以及為了支援這些「邊緣運算」所需的強大晶片（如 NVIDIA RTX 50 系列與各家 NPU）。此外，討論觸及了 FDA 新規範放寬後帶來的健康科技爆發，例如非侵入式血糖監測與智慧馬桶等。整體而言，CES 2026 標誌著科技正從數位的對話框，大步邁向能感知、移動並協助人類處理實體任務的時代。

VC Talks: Shao-Yu Jheng (Harvard) | Why Reindustrialization Won't Work Without Global Supply Networks

As reshoring and industrial policy return to the center of U.S. economic strategy, supply chains are increasingly framed as matters of national security and sovereignty. But history tells a more complicated story. In this episode of VC Talks, Uly revisits Left for Dead, a 1997 study on the revival of U.S. electronics, to question whether bringing manufacturing back home was ever the true driver of America’s industrial recovery.

Joined by Shao-Yu Jheng, a postdoctoral researcher in Economics at Harvard University, the conversation reframes the 1980s–90s comeback as a story of network leadership rather than national production. The episode examines why the Japanese manufacturing model ran into structural limits, how Taiwan and overseas Chinese (OC) production networks enabled coordination at scale, and why control over global supply networks—rather than factory location alone—proved decisive. Looking ahead, the discussion challenges today’s reshoring debate by asking what industrial policy can realistically change, and where its influence fundamentally ends.

在美國近年大力推動再工業化與製造回流的政策背景下，供應鏈被重新視為國家競爭力的核心資產。然而，歷史是否真的支持「把工廠搬回國內」就能重建產業實力？本集《VC Talks》回到一篇 1997 年探討美國電子產業復甦的研究〈Left for Dead〉，試圖從產業史的角度，重新檢視當前對 reshoring 的普遍想像。

主持人 Uly 與哈佛大學經濟學博士研究員鄭紹鈺深入對談，指出美國在 1980 至 1990 年代的產業反彈，並非來自製造回流，而是成功主導橫跨多國的生產網絡。節目聚焦日本製造模式的結構性限制，以及臺灣與海外華人（OC）產業網絡如何在關鍵時刻提供彈性、效率與協調能力，成為美國科技體系得以延續競爭優勢的隱形支柱。本集不只回顧歷史，也直指一個當代問題：在高度全球化的供應鏈體系中，真正的產業力量，究竟掌握在國家邊界，還是網絡控制之中？




📍 About Host: Uly Su
General Partner, Addins Venture
#CrossBorderMentor #StartupExpansion #TaiwanToSiliconValley
🌏 A cross-border mentor who supports startups in entering global markets, with deep experience in fundraising, talent strategy, and international operations across the U.S. and Taiwan.
🔗 LinkedIn: https://www.linkedin.com/in/ulysu/

📚 Learn more about Taiwan’s startup ecosystem 👉🏼 https://meetstartup.pse.is/4jhd5u

🚀 Stay connected with Taiwan–Silicon Valley innovation via the Startup Island TAIWAN Silicon Valley Hub:
https://sv.startupisland.tw/

Robotics and AI are becoming core foundational capabilities of the physical economy. As AI extends from cloud-based models into hospitals, logistics centers, and a wide range of service environments, robots increasingly carry functions of perception, reasoning, and decision-making, directly translating into measurable gains in productivity, system reliability, and operational safety. Together, they are emerging as critical infrastructure supporting the operation of modern society.

Taiwan has established “Robotics + AI” as a key direction in its future industrial布局, responding to the long-term challenge of a highly concentrated industrial structure. While semiconductors remain a central pillar, their capital-intensive nature and limited spillover effects mean that the next phase of growth must come from a new industrial curve. The National Development Council has therefore been tasked with identifying scalable, repeatable, and extensible industry models—channeling Taiwan’s deep strengths in hardware, engineering, and systems integration into professional, service-oriented robotics, with priority deployment in healthcare, long-term care, logistics, and facility management where demand is clear.

Along this trajectory, the government’s role is evolving into that of an industry accelerator. By integrating R&D outputs and supporting the formation of new robotics companies, it is laying the groundwork for private capital and startups to enter, scale, and participate sustainably. This episode approaches the topic through the lens of industrial structure and policy logic, explaining why Taiwan has positioned Robotics + AI at the core of its next industrial phase, and how this strategy may shape the country’s economic future.

機器人與 AI 正成為實體經濟運作的核心底層能力。隨著 AI 從雲端模型延伸至醫院、物流中心與各類服務場域，機器人逐步承載感知、判斷與決策功能，並直接轉化為可量化的生產力、系統穩定度與營運安全，成為支撐現代社會運作的重要基礎設施。

臺灣已將「機器人＋AI」確立為未來產業布局的關鍵方向，回應產業結構高度集中的長期挑戰。半導體仍是核心支柱，但其資本密集與產業外溢有限，使得下一階段成長動能必須來自新的產業曲線。國發會因此被賦予尋找可延展、可複製、可擴散產業模式的任務，將臺灣深厚的硬體、工程與系統整合能力，導入專業用、服務型機器人領域，並優先落地於醫療、長照、物流與場域管理等需求明確的應用場景。

在這個方向上，政府的角色將成為產業成形的推進器。透過整合研發成果整合、成立新的機器人企業，將可為民間資本與新創團隊鋪設可進場、可放大的基礎。本集從產業結構與政策邏輯切入，解析臺灣以機器人與 AI 作為下一階段產業核心的理由，以及這項布局對臺灣經濟可能帶來的深層影響。