<p>當 ChatGPT 於 2022 年 11 月 30 日向更廣泛的公眾發佈時，<strong>英偉達</strong> 的市值僅為 3450 億美元。截至 2025 年 7 月 25 日收盤，其市值已超過 4.2 萬億美元，使其成為全球最有價值的公司，且差距相當大。</p>
<p>鑑於這些歷史性的增長，對於許多成長型投資者來説，人工智能（AI）運動圍繞着英偉達展開並不令人驚訝。此時，該公司基本上被視為衡量整個 AI 領域整體健康狀況的晴雨表。</p>
<p>很難對英偉達下注，但我確實看到半導體領域中還有另一隻股票似乎在長期增長方面更具優勢。</p>
<p>讓我們探討一下為什麼 <strong>台積電</strong>（TSM 0.66%）在當前芯片領域是一個如此引人注目的機會，以及可能推動其回報超過英偉達的催化因素。</p>
<h2>AI 基礎設施浪潮剛剛開始</h2>
<p>我喜歡將 AI 敍事視為一個故事。在過去的幾年中，最大的篇章圍繞着被稱為圖形處理單元（GPU）的先進芯片組展開，這些芯片組被廣泛應用於各種生成性 AI 應用中。這些應用包括構建大型語言模型（LLM）、機器學習、機器人技術、自動駕駛汽車等。</p>
<p>這些不同的應用現在才開始逐漸顯現。AI 故事中的下一個大篇章是基礎設施將如何在實際開發和擴展這些更先進的技術中發揮作用。</p>
<p>全球管理諮詢公司麥肯錫公司估計，未來五年對 AI 基礎設施的投資可能達到 6.7 萬億美元，其中相當一部分將用於數據中心的硬件。</p>
<p>基於這一想法，考慮到雲計算巨頭 <strong>亞馬遜</strong>、<strong>微軟</strong> 和 <strong>谷歌</strong>，以及它們的 “七大巨頭” 同行 <strong>Meta Platforms</strong>，預計僅在今年就將投入超過 3300 億美元用於資本支出（capex）。其中大部分將用於額外的服務器、芯片和網絡設備，以加速 AI 數據中心的擴展。</p>
<p>在我看來，全球最大企業在資本支出上的激進投資是 AI 基礎設施浪潮開始成形的強烈信號。</p>
<p><img src="https://imageproxy.pbkrs.com/https://g.foolcdn.com/image//query-b3A9cmVzaXplJnVybD1odHRwczovL2cuZm9vbGNkbi5jb20vZWRpdG9yaWFsL2ltYWdlcy84MjY1NDEvZ2V0dHlpbWFnZXMtMTQ2ODI2NjE1MS0zLmpwZyZ3PTcwMA?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" width="700" height="393" original-src="https://imageproxy.pbkrs.com/https://g.foolcdn.com/image//query-b3A9cmVzaXplJnVybD1odHRwczovL2cuZm9vbGNkbi5jb20vZWRpdG9yaWFsL2ltYWdlcy84MjY1NDEvZ2V0dHlpbWFnZXMtMTQ2ODI2NjE1MS0zLmpwZyZ3PTcwMA"/></p>
<p>圖片來源：Getty Images。</p>
<h2>這對英偉達來説是個好消息，對台積電來説更是好消息</h2>
<p>不斷增長的 AI 基礎設施投資對英偉達來説是一個良好的順風，但對 <strong>超微</strong>、<strong>博通</strong> 和許多其他公司來説也是一個增長來源。</p>
<p>然而，與 AMD 或英偉達不同，台積電（簡稱 TSMC）的增長並不真正依賴於某一特定產品線的成功。換句話説，英偉達和 AMD 正在激烈競爭，以贏得 AI 工作負載，這歸結為它們誰能設計出最強大、最節能且價格合理的芯片。</p>
<p>圍繞台積電的投資案例在於，它可以被視為 AI 芯片市場中的一個更中立的參與者，因為其代工和製造服務將受益於推動 AI 基礎設施的更廣泛、更加世俗的順風——無論是哪個公司的芯片創造了最多的需求。</p>
<p>可以把台積電看作是實際製造出英偉達、AMD 和其他芯片公司所需的工具和設備的公司，這些公司需要在彼此之間競爭並進行銷售。</p>
<h2>台積電的 “英偉達時刻” 可能已經到來</h2>
<p>英偉達與台積電之間的估值差異表明了後者在更廣泛的芯片市場中的看法。</p>
<p><img src="https://imageproxy.pbkrs.com/https://media.ycharts.com/charts/d0ef808d5e9dc77b4a45d71e067bca97.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" width="720" height="456" original-src="https://imageproxy.pbkrs.com/https://media.ycharts.com/charts/d0ef808d5e9dc77b4a45d71e067bca97.png"/></p>
<p>TSMC 市盈率（前瞻性）數據來自 YCharts；市盈率 = 價格與收益之比。</p>
<p>像英偉達和 AMD 這樣的公司在很大程度上依賴於台積電的代工和製造服務，這些服務本質上是芯片行業的支柱。儘管華爾街的一些人會認為英偉達憑藉其芯片和軟件的雙重優勢擁有技術護城河，但我認為台積電擁有一個被低估的護城河，使其在芯片行業中相較於同行具有更廣泛的曝光度。</p>
<p>在未來五年中，隨着企業尋求超越當前的雲計算、網絡安全、企業軟件等市場，我認為 AI 開發的應用場景將會增加。</p>
<p>自動駕駛和量子計算等新興應用將進一步推動對 GPU 和數據中心容量的需求。因此，台積電可能正處於一個 “英偉達時刻” 的邊緣，迎來持續的爆炸性增長。</p>
<p>台積電的前瞻性市盈率（P/E）為 25，使其在與英偉達（前瞻性市盈率為 40）相比時處於獨特的位置，具有相當大的擴展潛力，因為 AI 的基礎設施篇章仍在繼續書寫。</p>
<p>我認為台積電的估值將在未來幾年內越來越與公司的增長相一致，因此我預測到 2030 年該股票將超過英偉達。</p>

台積電

半導體 3 倍做多 - Direxion

<p>一項預測表明，到 2030 年，台灣半導體制造公司（TSMC）將超過英偉達（Nvidia）的表現。儘管英偉達經歷了顯著增長，但 TSMC 在人工智能基礎設施市場中佔據關鍵地位，受益於對人工智能硬件日益增長的投資。TSMC 在芯片市場的中立角色，加上其市盈率低於英偉達，表明其有潛在的顯著增長空間。隨着人工智能應用的擴展，預計 TSMC 的估值將與其增長更加緊密地對齊，從而在未來幾年內實現優於英偉達的表現</p>

<p>When ChatGPT was released to the broader public on Nov. 30, 2022, <strong>Nvidia</strong> had a market capitalization of just $345 billion. As of the closing bell on July 25, 2025, its market cap had eclipsed $4.2 trillion, making it the most valuable company in the world -- by a pretty wide margin, too.</p>
<div class="lb-trans"><p>當 ChatGPT 於 2022 年 11 月 30 日向更廣泛的公眾發佈時，<strong>英偉達</strong> 的市值僅為 3450 億美元。截至 2025 年 7 月 25 日收盤，其市值已超過 4.2 萬億美元，使其成為全球最有價值的公司，且差距相當大。</p>
</div><p>Given these historic gains, it's not entirely surprising that for many growth investors, the artificial intelligence (AI) movement revolves around Nvidia. At this point, the company is basically seen as a barometer measuring the overall health of the entire AI sector.</p>
<div class="lb-trans"><p>鑑於這些歷史性的增長，對於許多成長型投資者來説，人工智能（AI）運動圍繞着英偉達展開並不令人驚訝。此時，該公司基本上被視為衡量整個 AI 領域整體健康狀況的晴雨表。</p>
</div><p>It's hard to bet against Nvidia, but I do see another stock in the semiconductor realm that appears better positioned for long-terms gains.</p>
<div class="lb-trans"><p>很難對英偉達下注，但我確實看到半導體領域中還有另一隻股票似乎在長期增長方面更具優勢。</p>
</div><p>Let's explore what makes <strong>Taiwan Semiconductor Manufacturing</strong>(TSM 0.66%) such a compelling opportunity in the chip space right now, and the catalysts at play that could fuel returns superior to Nvidia's over the next several years.</p>
<div class="lb-trans"><p>讓我們探討一下為什麼 <strong>台積電</strong>（TSM 0.66%）在當前芯片領域是一個如此引人注目的機會，以及可能推動其回報超過英偉達的催化因素。</p>
</div><h2>The AI infrastructure wave is just starting</h2>
<div class="lb-trans"><h2>AI 基礎設施浪潮剛剛開始</h2>
</div><p>I like to think of the AI narrative as a story. For the last few years, the biggest chapter revolved around advanced chipsets called graphics processing units (GPUs), which are used across a variety of generative AI applications. These include building large language models (LLMs), machine learning, robotics, self-driving cars, and more.</p>
<div class="lb-trans"><p>我喜歡將 AI 敍事視為一個故事。在過去的幾年中，最大的篇章圍繞着被稱為圖形處理單元（GPU）的先進芯片組展開，這些芯片組被廣泛應用於各種生成性 AI 應用中。這些應用包括構建大型語言模型（LLM）、機器學習、機器人技術、自動駕駛汽車等。</p>
</div><p>These various applications are only now beginning to come into focus. The next big chapter in the AI storyline is how infrastructure is going to play a role in actually developing and scaling up these more advanced technologies.</p>
<div class="lb-trans"><p>這些不同的應用現在才開始逐漸顯現。AI 故事中的下一個大篇章是基礎設施將如何在實際開發和擴展這些更先進的技術中發揮作用。</p>
</div><p>Global management consulting firm McKinsey &amp; Company estimates that investments in AI infrastructure could reach $6.7 trillion over the next five years, with a good portion of that allocated toward hardware for data centers.</p>
<div class="lb-trans"><p>全球管理諮詢公司麥肯錫公司估計，未來五年對 AI 基礎設施的投資可能達到 6.7 萬億美元，其中相當一部分將用於數據中心的硬件。</p>
</div><p>Piggybacking off of this idea, consider that cloud hyperscalers <strong>Amazon</strong>, <strong>Microsoft</strong>, and <strong>Alphabet</strong>, along with their &#34;Magnificent Seven&#34; peer <strong>Meta Platforms</strong>, are expected to devote north of $330 billion on capital expenditures (capex) just this year. Much of this is going toward additional servers, chips, and networking equipment for accelerated AI data center expansion.</p>
<div class="lb-trans"><p>基於這一想法，考慮到雲計算巨頭 <strong>亞馬遜</strong>、<strong>微軟</strong> 和 <strong>谷歌</strong>，以及它們的 “七大巨頭” 同行 <strong>Meta Platforms</strong>，預計僅在今年就將投入超過 3300 億美元用於資本支出（capex）。其中大部分將用於額外的服務器、芯片和網絡設備，以加速 AI 數據中心的擴展。</p>
</div><p>To me, the aggressive spending on capex from the world's largest businesses is a strong signal that the infrastructure wave in AI is beginning to take shape.</p>
<div class="lb-trans"><p>在我看來，全球最大企業在資本支出上的激進投資是 AI 基礎設施浪潮開始成形的強烈信號。</p>
</div><p><img src="https://imageproxy.pbkrs.com/https://g.foolcdn.com/image//query-b3A9cmVzaXplJnVybD1odHRwczovL2cuZm9vbGNkbi5jb20vZWRpdG9yaWFsL2ltYWdlcy84MjY1NDEvZ2V0dHlpbWFnZXMtMTQ2ODI2NjE1MS0zLmpwZyZ3PTcwMA" alt="" original-src="https://g.foolcdn.com/image/?url=https%3A%2F%2Fg.foolcdn.com%2Feditorial%2Fimages%2F826541%2Fgettyimages-1468266151-3.jpg&amp;op=resize&amp;w=700"/></p>
<p>Image Source: Getty Images.</p>
<div class="lb-trans"><p>圖片來源：Getty Images。</p>
</div><h2>This is great for Nvidia and even better for TSMC</h2>
<div class="lb-trans"><h2>這對英偉達來説是個好消息，對台積電來説更是好消息</h2>
</div><p>Rising AI infrastructure investment is a great tailwind for Nvidia but also a source of growth for <strong>Advanced Micro Devices</strong>, <strong>Broadcom</strong>, and many others.</p>
<div class="lb-trans"><p>不斷增長的 AI 基礎設施投資對英偉達來説是一個良好的順風，但對 <strong>超微</strong>、<strong>博通</strong> 和許多其他公司來説也是一個增長來源。</p>
</div><p>Unlike AMD or Nvidia, though, growth for Taiwan Semiconductor (TSMC for short) doesn't really hinge on the success of a particular product line. In other words, Nvidia and AMD are competing fiercely against one another to win AI workloads, which boils down to which of them can design the most powerful, energy efficient chips at an affordable price.</p>
<div class="lb-trans"><p>然而，與 AMD 或英偉達不同，台積電（簡稱 TSMC）的增長並不真正依賴於某一特定產品線的成功。換句話説，英偉達和 AMD 正在激烈競爭，以贏得 AI 工作負載，這歸結為它們誰能設計出最強大、最節能且價格合理的芯片。</p>
</div><p>The investment case around TSMC is that it could be seen as more of an agnostic player in the AI chip market because its foundry and fabrication services stand to benefit from broader, more secular tailwinds fueling AI infrastructure -- regardless of whose chips create the most demand.</p>
<div class="lb-trans"><p>圍繞台積電的投資案例在於，它可以被視為 AI 芯片市場中的一個更中立的參與者，因為其代工和製造服務將受益於推動 AI 基礎設施的更廣泛、更加世俗的順風——無論是哪個公司的芯片創造了最多的需求。</p>
</div><p>Think of TSMC as the company actually making the picks and shovels that Nvidia, AMD, and other chip companies need to go out and sell while competing among one another.</p>
<div class="lb-trans"><p>可以把台積電看作是實際製造出英偉達、AMD 和其他芯片公司所需的工具和設備的公司，這些公司需要在彼此之間競爭並進行銷售。</p>
</div><h2>TSMC's &#34;Nvidia moment&#34; may be here</h2>
<div class="lb-trans"><h2>台積電的 “英偉達時刻” 可能已經到來</h2>
</div><p>The valuation disparity between Nvidia and TSMC says a couple of things about how the latter is viewed in the broader chip landscape.</p>
<div class="lb-trans"><p>英偉達與台積電之間的估值差異表明了後者在更廣泛的芯片市場中的看法。</p>
</div><p><img src="https://imageproxy.pbkrs.com/https://media.ycharts.com/charts/d0ef808d5e9dc77b4a45d71e067bca97.png" alt="" original-src="https://media.ycharts.com/charts/d0ef808d5e9dc77b4a45d71e067bca97.png"/></p>
<p>TSM PE Ratio (Forward) data by YCharts; PE = price to earnings.</p>
<div class="lb-trans"><p>TSMC 市盈率（前瞻性）數據來自 YCharts；市盈率 = 價格與收益之比。</p>
</div><p>Companies such as Nvidia and AMD rely heavily on TSMC's foundry and fabrication services, which are essentially the backbone of the chip industry. While some on Wall Street would argue that Nvidia has a technological moat thanks to its one-two punch of chips and software, I think that TSMC has an underappreciated moat that provides the company with broader exposure to the chip industry compared to its peers.</p>
<div class="lb-trans"><p>像英偉達和 AMD 這樣的公司在很大程度上依賴於台積電的代工和製造服務，這些服務本質上是芯片行業的支柱。儘管華爾街的一些人會認為英偉達憑藉其芯片和軟件的雙重優勢擁有技術護城河，但我認為台積電擁有一個被低估的護城河，使其在芯片行業中相較於同行具有更廣泛的曝光度。</p>
</div><p>Over the next five years, I think use cases of AI development increase as businesses seek to expand beyond their current markets in cloud computing, cybersecurity, enterprise software, among others.</p>
<div class="lb-trans"><p>在未來五年中，隨着企業尋求超越當前的雲計算、網絡安全、企業軟件等市場，我認為 AI 開發的應用場景將會增加。</p>
</div><p>Emerging applications such as autonomous driving and quantum computing will drive demand for GPUs and data center capacity even further. For this reason, TSMC may be on the verge of an &#34;Nvidia moment&#34; featuring prolonged, explosive growth.</p>
<div class="lb-trans"><p>自動駕駛和量子計算等新興應用將進一步推動對 GPU 和數據中心容量的需求。因此，台積電可能正處於一個 “英偉達時刻” 的邊緣，迎來持續的爆炸性增長。</p>
</div><p>TSMC's modest forward price-to-earnings multiple (P/E) of 25 puts it in a unique position compared to Nvidia (with its forward P/E of 40) for considerable expansion as the infrastructure chapter of AI continues to be written.</p>
<div class="lb-trans"><p>台積電的前瞻性市盈率（P/E）為 25，使其在與英偉達（前瞻性市盈率為 40）相比時處於獨特的位置，具有相當大的擴展潛力，因為 AI 的基礎設施篇章仍在繼續書寫。</p>
</div><p>I think Taiwan Semiconductor Manufacturing's valuation will increasingly become more congruent with the company's growth over the next several years, and so I predict that the stock will outperform Nvidia by 2030.</p>
<div class="lb-trans"><p>我認為台積電的估值將在未來幾年內越來越與公司的增長相一致，因此我預測到 2030 年該股票將超過英偉達。</p>
</div>

Prediction: This Artificial Intelligence (AI) Stock Will Outperform Nvidia by 2030

- Nvidia 市值從 2022 年的 3450 億美元增至 2025 年的 4.2 萬億美元。  
- TSMC 被視為 AI 基礎設施投資的受益者，預計將實現顯著增長。  
- 未來幾年，TSMC 的估值有望與其增長相匹配，可能超越 Nvidia。  

預測：這隻人工智能（AI）股票將在 2030 年之前表現優於英偉達