Taiwan Semiconductor Manufacturing Co.(NYSE:TSM) is betting that the next wave of artificial intelligence won't just require more computing power — it will require machines that can see and understand the physical world.
台積電（NYSE:TSM）認為，人工智慧的下一波浪潮不僅需要更強的計算能力，還需要能夠感知並理解物理世界的機器。
TSMC andSony Group Corp(NYSE:SONY) are planning to invest roughly $6.3 billion in a joint venture to produce next-generation image sensors in Japan, with commercial production potentially beginning as early as 2029. Sony will own 60% of the venture and TSMC 40%. The companies also plan to explore physical-AI applications including automotive and robotics.
台積電和索尼集團（NYSE:SONY）計劃共同投資約$63億，在日本成立一家合資企業，生產下一代圖像傳感器，商業生產最早可能於2029年開始。索尼將持有該合資企業60%的股份，台積電持股40%。雙方還計劃探索包括汽車和機器人在內的實體人工智慧（physical-AI）應用。
The first wave of AI was built around a simple idea: Give computers more computing power.Nvidia Corp.(NASDAQ) supplied much of the hardware that became the industry's AI "brain."
第一波人工智慧基於一個簡單的理念：為計算機提供更多的計算能力。輝達公司（納斯達克）提供了大量硬體，成為行業人工智慧的「大腦」。
TSMC isn't walking away from that booming business making processors for Nvidia and other AI-chip designers. Instead, its push into image sensors could give the chipmaker exposure to what comes next: AI systems that don't just compute, but see, interpret and interact with the world around them.
台積電並未放棄為輝達及其他AI晶片設計公司製造處理器這一蓬勃發展的業務。相反，其進軍圖像傳感器領域，或將使這家晶片製造商接觸到下一個階段：不僅能計算，還能感知、理解並與周圍世界互動的人工智慧系統。
AI Is Moving Beyond the Data Center
人工智慧正超越數據中心
Today's AI boom is largely powered by data centers packed with specialized processors. Nvidia's GPUs provide the computing power needed to train and run increasingly sophisticated AI models.
當前的人工智慧熱潮主要由配備專用處理器的數據中心驅動。輝達的GPU提供了訓練和運行日益複雜的人工智慧模型所需的計算能力。
But robots and autonomous vehicles need another capability before they can act on that intelligence: they need to perceive their surroundings.
但機器人和自動駕駛汽車在據此採取行動之前，還需要另一種能力：它們需要感知周圍環境。
Image sensors convert what a camera sees into electronic information that machines can process. Sony is already a global leader in image sensors used in smartphones and vehicles, while TSMC brings its manufacturing scale and advanced semiconductor processes to the partnership.
圖像傳感器將攝像頭所見內容轉換為機器可處理的電子資訊。索尼已是全球智能手機和汽車所用圖像傳感器的領軍企業，而台積電則為此次合作帶來了其製造規模和先進的半導體工藝。
That creates a potentially important new layer of the AI hardware market.
這可能在人工智慧硬體市場中形成一個重要的新層級。
A robot needs computing power to make decisions. It also needs sensors to understand where it is, what is around it and what is changing.
機器人需要計算能力來做出決策，同時也需要傳感器來理解自身位置、周圍環境以及正在發生的變化。
Read Also:Seagate Says Physical AI Will Be Built on Video, Not Robots
另請閱讀：希捷稱實體AI將基於視頻而非機器人構建
TSMC's Timing Is Significant
台積電的時機選擇意義重大
TSMC is already expanding aggressively to meet demand for AI computing.
台積電已在積極擴張，以滿足AI計算需求。
The company raised its 2026 capital budget to $60 billion-$64 billion, with 70%-80% earmarked for advanced process technologies and another 10%-20% for advanced packaging, testing, mask making and related areas. TSMC also expects full-year revenue growth of more than 40%. That investment is primarily supporting the existing AI and high-performance-computing boom.
該公司將其2026年資本支出預算上調至$600億-$640億，其中70%–80%用於先進製程技術，另有10%–20%用於先進封裝、測試、光罩製造及相關領域。台積電還預計全年營收增長將超過40%。這項投資主要支援當前AI和高性能計算的熱潮。
The Sony partnership points toward a broader opportunity.
與索尼的合作指向更廣闊的機會。
If physical AI takes off across robots, autonomous vehicles and industrial machines, semiconductor content could expand beyond the processors running AI models to the sensors helping those machines understand the physical world.
如果實體AI在機器人、自動駕駛汽車和工業設備等領域興起，半導體的應用範圍可能不僅限於運行AI模型的處理器，還將擴展到幫助這些設備感知物理世界的傳感器。
That does not make Sony's image-sensor project a guaranteed new growth engine for TSMC. Production isn't expected to begin until 2029, and the companies have not disclosed how much revenue the venture could eventually generate. But it gives investors another way to think about TSMC's AI exposure.
但這並不意味著索尼的圖像傳感器項目必然會成為台積電新的增長引擎。預計生產要到2029年才會啓動，且雙方尚未披露該合資項目最終可能帶來多少收入。不過，這為投資者提供了另一種看待台積電AI業務敞口的方式。
Nvidia is helping build the brains of the AI economy. TSMC is already manufacturing those brains — and its Sony partnership suggests it may also want a role in building the technology that helps AI see.
輝達正在幫助構建AI經濟的「大腦」。台積電已經在製造這些「大腦」——而其與索尼的合作表明，它或許也希望在幫助AI「看見」的技術領域佔據一席之地。 (Benzinga)