<p>在我們即將結束 2025 年之際，人工智能泡沫幾乎毫髮無損地度過了這一年。不過，關於估值的擔憂逐漸上升。在談論這個問題時，可以公平地説，這一切將在 2026 年的討論中佔據重要地位。那麼問題是，市場是否對人工智能的影響過於樂觀？明年我們會看到現實的檢驗嗎？</p>
<p>嗯，這確實是值得思考和考慮的事情。</p>
<p>對人工智能的簡單理解是，它通過提高流程的效率和速度來提升生產力。讓我們以榨橙汁為例。是的，我喜歡用水果作為例子。這總是讓我想起這篇文章，解釋當年的 LIBOR 醜聞。</p>
<p>但沒錯，榨橙汁。</p>
<p>假設你是一個榨橙汁的人，某天你宣佈你將購買一台高科技、超快速的橙子去皮和榨汁機，以便準備好榨汁出售。人們對此感到興奮，儘管你當時的利潤只有 5 美元，他們還是給了你 500 美元。</p>
<p>人們並不在乎今天的錢，因為他們 “相信” 有了新技術，你將徹底改變榨橙汁的銷售世界。</p>
<p>所以，這基本上就是我們在整個人工智能泡沫中所處的狀態。現實的檢驗尚未發生，但只需時間問題，以下問題就會被提出：</p>
<ul>
<li>新技術真的那麼好嗎？</li>
<li>它真的提高了榨汁準備的效率和時間成本嗎？</li>
<li>它真的幫助大幅提高了利潤率嗎？</li>
</ul>
<p>如果將這些問題轉化為那些深陷人工智能投資的公司和企業，這些都是在某個時刻有效的問題。這可能是投資者明年所要求的。</p>
<p>在此之前，市場會為人工智能投資和增加資本支出而歡呼，認為這是革命性的。現在，這樣做並不新鮮，而是相對 “普遍”。</p>
<p>就像新款 PlayStation 5 發佈時一樣。你是那個酷孩子，大家都想和你一起玩。但當其他人也開始擁有它時，你所擁有的就沒有什麼不同，人們反而更願意待在自己家裏。</p>
<p>因此，問題轉向如何讓人們，即投資者，留下來？是什麼讓你的產品更 “神奇” 和 “特別”？這就是生產力討論的所在。</p>
<p>對於大型科技公司來説，討論不再是關於在人工智能上的支出，而是關於誰能真正正確地使用這些技術以反映更好的底線。</p>
<p>對於谷歌和 Meta 來説，關鍵在於將其轉化為廣告收入，前者還將受到對其雲業務的審查。到目前為止，他們是兩家更容易展示如何提高生產力以及如何將其轉化為整體收益的公司。</p>
<p>然後是亞馬遜和微軟，他們都投入了大量資本，試圖説服投資者他們在人工智能領域保持競爭力。</p>
<p>現在，亞馬遜在人工智能上的資本支出是所有公司中最多的，他們目前隱藏在背後的一個理由是，他們的收入流和生產力提升分佈在多個方面。他們有倉庫技術、機器人、網站和雲系統，所有這些都與人工智能的進步相結合。因此，利潤必須不斷湧入，以説服投資者抵消他們的大量支出。</p>
<p>也就是説，亞馬遜足夠大，可以使自己免受依賴芯片製造商和外部數據中心的風險。他們確實在努力開發自己的芯片，並在後者上大力擴展。我上週在這裏談到了數據中心和對電力爭奪的重要性。</p>
<p>至於微軟，情況相對簡單，Copilot 是他們最大的推銷產品。證明在於數字，即有多少人真正感到需要註冊該公司提供的人工智能軟件。就我個人而言，我並不是很喜歡，自己的偏好是繼續使用 Windows 10。</p>
<p>我們不能談論大型科技公司而不提到整個人工智能泡沫的代言人，對吧？在這次熱潮中，Nvidia 是最大的名字，現在是時候看看這些高期望是否終於會追上他們了？</p>
<p>Blackwell 芯片的發佈顯示，需求仍然遠遠超過供應。但如果積壓開始減少，像亞馬遜和微軟這樣的公司也開始開發自己的人工智能生態系統，這可能對 Nvidia 來説是一個令人擔憂的信號，因為他們面臨着不斷超越和交付超出預期的壓力。</p>
<p>別誤會，Nvidia 仍然是一個主要的現金奶牛，也是人工智能泡沫中最大的收益者。但投資者的期望是否過高，以至於任何不完美的表現都會受到懲罰？這將是一個有趣的觀察，尤其是來自中國市場的關鍵風險可能會帶來一些不合時宜的頭條新聞。</p>
<p>但如果一切順利，Jensen Huang 和他的公司可能會成為首個市值達到 5 萬億美元的股票。或者如果你想大夢一場，甚至可能達到 10 萬億美元。</p>

META

<p>這篇文章探討了 2026 年大型科技公司的潛在重置，質疑市場是否對人工智能的影響過於樂觀。文章強調，像谷歌、Meta、亞馬遜、微軟和英偉達這樣的公司需要展示出來自人工智能投資的實際生產力提升和利潤改善，暗示現實檢查可能即將到來</p>

<p>As we look to wrap up 2025, the AI bubble just about managed to get away unscathed to end the year. That being said, there were rising concerns to deal with especially that on valuation. And in talking about that, it is fair to say that all of this will be a mainstay in the conversation for 2026. So the question is, have markets gotten too optimistic about the impact of AI? And are we going to see a reality check come next year?</p>
<div class="lb-trans"><p>在我們即將結束 2025 年之際，人工智能泡沫幾乎毫髮無損地度過了這一年。不過，關於估值的擔憂逐漸上升。在談論這個問題時，可以公平地説，這一切將在 2026 年的討論中佔據重要地位。那麼問題是，市場是否對人工智能的影響過於樂觀？明年我們會看到現實的檢驗嗎？</p>
</div><p>Well, it definitely is something worth thinking about and considering.</p>
<div class="lb-trans"><p>嗯，這確實是值得思考和考慮的事情。</p>
</div><p>The simple understanding of AI is that it boosts productivity by making processes more efficient and faster right. Let's take an intelligible example of making orange juice from the fruit itself. Yes, I love fruit examples. It always brings me back to this article here in explaining the whole LIBOR scandal back in the day.</p>
<div class="lb-trans"><p>對人工智能的簡單理解是，它通過提高流程的效率和速度來提升生產力。讓我們以榨橙汁為例。是的，我喜歡用水果作為例子。這總是讓我想起這篇文章，解釋當年的 LIBOR 醜聞。</p>
</div><p>But yes, orange juice.</p>
<div class="lb-trans"><p>但沒錯，榨橙汁。</p>
</div><p>Let's say you are someone who squeezes orange juice to sell, and one day you make it known that you are going to buy a high-tech and super-quick orange peeler and squeezer to get the juice ready to sell. People get excited about that and throw you $500 even though you only make like $5 in profits at the time.</p>
<div class="lb-trans"><p>假設你是一個榨橙汁的人，某天你宣佈你將購買一台高科技、超快速的橙子去皮和榨汁機，以便準備好榨汁出售。人們對此感到興奮，儘管你當時的利潤只有 5 美元，他們還是給了你 500 美元。</p>
</div><p>The people aren't fussed about the money today because they &#34;believe&#34; that with the new technology, you're going to revolutionise the world of selling orange juice.</p>
<div class="lb-trans"><p>人們並不在乎今天的錢，因為他們 “相信” 有了新技術，你將徹底改變榨橙汁的銷售世界。</p>
</div><p>So, that's pretty much where we were or somewhat still are at in the whole AI bubble. The sense check hasn't quite happened yet but it's only a matter of time until questions are asked about the following:</p>
<div class="lb-trans"><p>所以，這基本上就是我們在整個人工智能泡沫中所處的狀態。現實的檢驗尚未發生，但只需時間問題，以下問題就會被提出：</p>
</div><ul>
<li>Is the new technology really <em>that</em> good?</li>
<li>How has it really improved the efficiency and time cost of getting the orange juice ready for sale?</li>
<li>Has it really helped to increase profit margin by a great amount?</li>
</ul>
<div class="lb-trans"><ul>
<li>新技術真的那麼好嗎？</li>
<li>它真的提高了榨汁準備的效率和時間成本嗎？</li>
<li>它真的幫助大幅提高了利潤率嗎？</li>
</ul>
</div><p>If you translate that to companies and firms that are knee deep in AI investment, these are all valid questions at some point. And that could be what investors are demanding next year.</p>
<div class="lb-trans"><p>如果將這些問題轉化為那些深陷人工智能投資的公司和企業，這些都是在某個時刻有效的問題。這可能是投資者明年所要求的。</p>
</div><p>Before this, markets would cheer on AI investment and increased capital expenditure to be <em>revolutionary</em>. Now, doing so isn't anything new but instead it's rather <em>commonplace</em> instead.</p>
<div class="lb-trans"><p>在此之前，市場會為人工智能投資和增加資本支出而歡呼，認為這是革命性的。現在，這樣做並不新鮮，而是相對 “普遍”。</p>
</div><p>It's like having the new PlayStation 5 on release. You're the cool kid and everyone wants to hang out with you when you have it. But then when everyone else also starts to own it, what you have isn't anything different and people hang out at their own homes instead.</p>
<div class="lb-trans"><p>就像新款 PlayStation 5 發佈時一樣。你是那個酷孩子，大家都想和你一起玩。但當其他人也開始擁有它時，你所擁有的就沒有什麼不同，人們反而更願意待在自己家裏。</p>
</div><p>And so the question then turns to how do you get the people i.e. investors to stay? What makes yours more &#34;magical&#34; and &#34;special&#34;? That is where the productivity conversation comes in.</p>
<div class="lb-trans"><p>因此，問題轉向如何讓人們，即投資者，留下來？是什麼讓你的產品更 “神奇” 和 “特別”？這就是生產力討論的所在。</p>
</div><p>For Big Tech, that means the conversation isn't anymore about spending on AI. It's about who can actually use that correctly to reflect a better bottom line.</p>
<div class="lb-trans"><p>對於大型科技公司來説，討論不再是關於在人工智能上的支出，而是關於誰能真正正確地使用這些技術以反映更好的底線。</p>
</div><p>For the likes of Google and Meta, it's all about translating that to ad revenue with the former also going to be scrutinised on their cloud business. And so far, they are two of the better ones that have an easier time to show how increased productivity and how that translates to earnings in general.</p>
<div class="lb-trans"><p>對於谷歌和 Meta 來説，關鍵在於將其轉化為廣告收入，前者還將受到對其雲業務的審查。到目前為止，他們是兩家更容易展示如何提高生產力以及如何將其轉化為整體收益的公司。</p>
</div><p>Then you have the likes of Amazon and Microsoft, who both have laid out massive amounts of capital in trying to convince investors that they are keeping up in the AI game.</p>
<div class="lb-trans"><p>然後是亞馬遜和微軟，他們都投入了大量資本，試圖説服投資者他們在人工智能領域保持競爭力。</p>
</div><p>Now, Amazon has committed the most in terms of capital expenditure on AI as compared to everyone else and one thing they are hiding behind for now is that their revenue stream and productivity gains are spread across multiple points. They have their warehouse technologies, robots, website, and cloud systems all layered with AI advancements. And so, the profits have to keep rolling in to convince investors against their big amount of money spent.</p>
<div class="lb-trans"><p>現在，亞馬遜在人工智能上的資本支出是所有公司中最多的，他們目前隱藏在背後的一個理由是，他們的收入流和生產力提升分佈在多個方面。他們有倉庫技術、機器人、網站和雲系統，所有這些都與人工智能的進步相結合。因此，利潤必須不斷湧入，以説服投資者抵消他們的大量支出。</p>
</div><p>That said, Amazon is also big enough to insulate themselves from risks of having to rely on chipmakers and external data centers. They do work to develop their own chips and are going big in expanding on the latter as well. I spoke about data centers and the importance of the fight for power last week here.</p>
<div class="lb-trans"><p>也就是説，亞馬遜足夠大，可以使自己免受依賴芯片製造商和外部數據中心的風險。他們確實在努力開發自己的芯片，並在後者上大力擴展。我上週在這裏談到了數據中心和對電力爭奪的重要性。</p>
</div><p>As for Microsoft, it's quite straightforward with Copilot being their biggest push product offering. The proof will be in the numbers, that being how many people actually feel the need to sign up for AI software delivered by the firm. And personally speaking, I'm not a big fan with my own taste preference being to continue using Windows 10.</p>
<div class="lb-trans"><p>至於微軟，情況相對簡單，Copilot 是他們最大的推銷產品。證明在於數字，即有多少人真正感到需要註冊該公司提供的人工智能軟件。就我個人而言，我並不是很喜歡，自己的偏好是繼續使用 Windows 10。</p>
</div><p>And we can't talk about Big Tech without talking about the poster boy of the whole AI bubble now, can we? Nvidia has been the biggest name of them all during this run and is it time that the lofty expectations finally catch up to them?</p>
<div class="lb-trans"><p>我們不能談論大型科技公司而不提到整個人工智能泡沫的代言人，對吧？在這次熱潮中，Nvidia 是最大的名字，現在是時候看看這些高期望是否終於會追上他們了？</p>
</div><p>The Blackwell chip release shows that demand is still well outweighing supply. But if backlogs start to reduce and companies like Amazon and Microsoft also start developing their own AI ecosystem, that could be a troubling sign for Nvidia amid the pressure to constantly outperform and deliver well above what they are doing.</p>
<div class="lb-trans"><p>Blackwell 芯片的發佈顯示，需求仍然遠遠超過供應。但如果積壓開始減少，像亞馬遜和微軟這樣的公司也開始開發自己的人工智能生態系統，這可能對 Nvidia 來説是一個令人擔憂的信號，因為他們面臨着不斷超越和交付超出預期的壓力。</p>
</div><p>Don't get me wrong. Nvidia is still a major cash cow and the biggest earner from the continued focus in the AI bubble. But are investor expectations too high that anything less than perfect will get punished? That will be interesting to see, especially with key risks from the China market that could provide some untimely headlines.</p>
<div class="lb-trans"><p>別誤會，Nvidia 仍然是一個主要的現金奶牛，也是人工智能泡沫中最大的收益者。但投資者的期望是否過高，以至於任何不完美的表現都會受到懲罰？這將是一個有趣的觀察，尤其是來自中國市場的關鍵風險可能會帶來一些不合時宜的頭條新聞。</p>
</div><p>But if all goes well for Jensen Huang and his company, they could be the first ever $5 trillion market cap stock. Or if you want to dream big, maybe even $10 trillion.</p>
<div class="lb-trans"><p>但如果一切順利，Jensen Huang 和他的公司可能會成為首個市值達到 5 萬億美元的股票。或者如果你想大夢一場，甚至可能達到 10 萬億美元。</p>
</div>

Will 2026 mark the big reset for Big Tech?

- 2025 年人工智能泡沫面臨估值擔憂，市場過於樂觀。  
- 大型科技公司需展示 AI 對業績的實際改善。  
- Nvidia 面臨高期望壓力，市場風險需關注。

2026 年是否會成為科技巨頭的大重置時刻？