Name: Nvidia 黃仁勳談論AI保麗龍、川普、以及與中國的軍備競賽！ (Nvidia's Jensen Huang on an AI Bubble, Trump, and the Arms Race with China)
Uploaded: 2026-01-16T10:08:26.000Z
Duration: 24 min 32 s

轉承詞

Everybody's jobs will change as a result of AI.

AI 的出現，會改變每個人的工作。

Obviously, every single industrial revolution, some jobs are just gone, but

顯然，每一次的工業革命，都會有一些工作消失，但是

Everybody will have to use AI, because if you don't use AI, you're going to lose your job to somebody who does.

每個人都必須使用 AI，因為如果你不用 AI，你的工作就會被用 AI 的人搶走。

Time has named the architects of AI as the 2025 Person of the Year.

Time 雜誌將 AI 的締造者們選為 2025 年度風雲人物。

One of those architects is Jensen Huang, the CEO of Nvidia, which is now the world's most valuable company.

其中一位締造者是 Nvidia 的執行長 Jensen Huang，Nvidia 現在是全球最有價值的公司。

Nvidia has a near monopoly on the supply of GPUs, the advanced computer chips powering the AI boom that's transforming the planet.

Nvidia 在 GPU 的供應上幾乎壟斷，GPU 是驅動這波 AI 浪潮、正在改變世界的先進電腦晶片。

The industry's rapid growth has tech leaders proclaiming a new era of abundance, while skeptics warn of a looming financial bubble.

這個產業的快速成長，讓科技巨頭們宣稱一個新的豐盛時代來臨，同時懷疑論者則警告即將到來的金融保麗龍。

In November, I went to speak with Huang at Nvidia's Bay Area headquarters.

去年十一月，我去 Nvidia 灣區總部跟 Huang 聊了一下。

Google CEO Sundar Pichai said this week that there were elements of irrationality in the stock market around AI right now.

Google 執行長 Sundar Pichai 這週說，現在股市裡有些跟 AI 相關的非理性成分。

There's always some irrationality somewhere in the stock market.

股市裡總是有一些非理性存在的。

I would say, though, the part that most people miss is that underneath the foundation of chatbots and agentic AI is a transformation of computing from general purpose to accelerated computing.

不過我會說，大部分人忽略的部分是，在哈拉機器人和代理式 AI 的基礎之下，運算正在從通用型轉變為加速運算。

When do you then then gauge the investments incremental on top of that relative to the opportunity of generative or agentic AI AGI on top of that?

那你怎麼衡量在生成式 AI 或代理式 AI AGI 之上的增量投資，相對於其機會呢？

Then you'll come to the conclusion, in fact, it's very modest.

那你就會得出結論，事實上，它相當溫和。

It's still significant, but it's modest relative to the overall computing market.

它仍然很重要，但相對於整體運算市場來說，算是溫和的。

So you're somewhere between 4 and $5 trillion company right now.

所以你現在大概是四到五兆美元的公司。

A lot of that cash flow that you're bringing in as revenue, you're investing in other companies in the space.

你們賺進來的現金流，大部分都拿去投資了這個領域的其他公司。

Some people in the world of analysts and the stock market are saying this creates a degree of circularity.

分析師和股市圈裡有些人認為，這會造成某種程度的循環。

這可能會產生額外的系統性風險。

If, for example, companies like OpenAI, which is not profitable, fails to meet the commitments that it's making to buy data centers mostly full of your chips.

舉例來說，如果像 OpenAI 這樣還沒賺錢的公司，無法履行它們對購買資料中心（裡面大部分是你們的晶片）的承諾。

Are you concerned that there is a growing structural risk related to all of those things?

你們會擔心，跟所有這些事情相關的結構性風險正在增加嗎？

Not because of this reason, for any other reason?

不是因為這個原因，而是因為其他原因嗎？

Well, there are other reasons that we can talk about, but not for this particular reason.

嗯，還有其他原因我們可以聊聊，但不是因為這個特定原因。

Every single transaction we do, every single purchase order that we take, we still go back to first principles and we ask ourselves, do you really have the demand for it?

我們做的每一筆交易、接的每一個採購訂單，我們都會回到最根本的原則，問自己：你們真的有這個需求嗎？

Because if you don't have the demand for it, somebody else has demand for it.

因為如果你們沒有這個需求，別人會有。

The reason why we invest is several reasons.

If you look at the investments that we will likely make and have made up to now, and

如果你看看我們過去到現在可能做的投資，

you compare it to the revenues that we're generating, it's in the low single digits.

再對比我們產生的營收，大概是個位數的低百分比。

And so that kind of tells you how could low single digits create circularity?

這就告訴你，這麼低的百分比要怎麼創造循環？

And then very importantly, this is the last reason.

然後非常重要的一點，這是最後一個原因。

These are consequential companies, most important companies of our time, and

這些都是影響深遠的公司，是我們這個時代最重要的公司，

we have the opportunity to have a front seat, front row seat partnership with them.

而我們有機會和他們建立第一線、最前排的合作夥伴關係。

We kind of also have, if you will, inside information about how successful they're doing.

我們也等於是，如果你們願意這麼說，能獲得他們成功資訊的內部消息。

And so the opportunity to invest in them.

It's going to turn out to be a great investment.

I'd love to move on to talking about the year that you've had in the world of product, politics and policy.

我很想接著聊聊您在產品、政治和政策領域的這一年。

The year begins with Trump entering the Oval Office.

這一年始於川普入主橢圓形辦公室。

I understand that your first meeting with him ever was that month in January.

我知道您第一次見到他就是在那個一月。

And since then you've traveled the world with him.

從那之後，您就和他一起走遍世界。

He's referred to you as a close friend and a great leader.

他稱您為摯友和偉大的領導者。

And I'm curious, having spent so much time with him up close this year, what's something that you've learned about his approach to AI and the way that he does business?

我很好奇，今年有這麼多近距離和他相處的時間，您從中學到了什麼關於他處理 AI 的方式，以及他做生意的方法？

I guess maybe I'll say some things that isn't evident watching or reading about President Trump on TV or news.

我猜或許我會說一些，從電視或新聞上看川普總統時，或閱讀關於他的報導時，看不出來的事情。

First of all, he's an extraordinarily good listener.

首先，他是一個非常非常棒的傾聽者。

He's almost, I think everything I've ever said to him, he's remembered.

他幾乎，我想我跟他講過的每一件事，他都記住了。

他是一個非常非常努力工作的人。

His work ethic, he starts as early as anybody else and oftentimes we'll be talking about things late in the night and I'm kind of ready for bed and he would have gone for several more hours.

他的工作態度，他跟別人一樣早開始，而且我們常常在深夜還在討論事情，我差不多準備要睡覺了，但他還可以再多做幾個小時。

And so I've never met anybody with such incredible work ethic.

所以我從沒見過像他這樣工作態度這麼驚人的人。

And so it just came from everything that he's done.

所以這一切都來自他所做的一切。

As far as I can tell, there is no after work spend time with doing something else.

就我所知，沒有所謂的下班後的時間去從事別的事情。

And so these two qualities are really impressive.

所以這兩種特質真的令人印象深刻。

And I wouldn't have guessed those two qualities, but

he's excellent listener, he's an incredible work ethic and he's super smart.

他非常善於傾聽，工作態度超認真，頭腦也很聰明。

In the time that you spent in and around the Trump administration and traveling the world this year, is there an important moment that you experienced that the world doesn't know about yet, that you think it should?

在你待在川普政府時期，以及今年周遊列國的這段時間裡，有沒有什麼你經歷過、但全世界還不知道、你覺得應該要知道的重要時刻？

The world is unaware of how deeply dependent United States and China are to each other.

全世界不知道美國和中國有多麼地相互依賴。

The ideas that floated around about United States decoupling from China, I think is flawed.

關於美國與中國脫鉤的想法，我覺得是有缺陷的。

And our dependency on each other is quite significant.

而我們之間的依賴關係其實相當重大。

And every single time we press it, we realize what else we're dependent on each other on.

每次我們去觸碰它，都會發現我們還在哪些方面相互依賴。

This will be the most consequential relationship in the next century.

這將是下個世紀最重要的關係。

If this is so, I think that we really ought to recognize that managing the relationship and not causing conflict is the most important thing we could do.

如果真是這樣，我認為我們真的應該認知到，處理好這段關係、避免衝突，是我們能做的最重要的事情。

And so negotiations, communications, compromise, and just the continuous process of doing that is essential.

所以協商、溝通、妥協，以及持續不斷地進行這些事情，是至關重要的。

And I think the current administration has demonstrated incredible wisdom in engaging, whether it's hexth with the military, whether it's Besant and Treasury, President Trump himself, Ambassador Greer, all of them representing us, engaging deeply in conversations there.

我認為現任政府在接觸上展現了非凡的智慧，無論是與軍方、財政部、川普總統本人，或是葛瑞爾大使，他們都代表我們，深入地進行對話。

And I think it'll demonstrate that a century from now that impactful change versus isolation and avoidance.

我認為這將證明，一個世紀後，有影響力的改變勝過孤立和迴避。

I think the second thing that I would say that the world didn't realize is how deeply dependent we are in the AI industry on the brilliant students and the brilliant scientists of China.

我認為第二點，我想說的是，全世界沒有意識到，在AI產業裡，我們對中國的優秀學生和頂尖科學家有多麼深的依賴。

50% of the world's AI researchers are from China.

全球50%的AI研究人員來自中國。

Their companies in China wants China to win.

And I think the Chinese want China to win.

而且我認為中國人也希望中國贏。

We want America to win and we can have a healthy competition while we compete, compete fairly and collaborate at the same time.

我們希望美國贏，我們可以在競爭的同時，公平競爭，同時合作。

But I think the taking it emotionally too far from that results in consequences in relationships that are just harder to manage.

但我認為，過度情緒化地看待這件事，會導致關係中出現更難處理的後果。

You've been still unable to ship, at the beginning of the administration, your H20 chip to China and had argued successfully ultimately to the Trump administration for them to remove that export control.

在政府初期，你們仍然無法將H20晶片運往中國，並成功爭取川普政府取消了出口管制。

However, shortly after that, Commerce Secretary Howard Lutnick said that the goal of American chip policy in that regard was to get China addicted to US Chips.

然而，不久之後，商務部長霍華德·魯特尼克表示，美國在這方面的晶片政策目標是讓中國對美國晶片上癮。

And after that, China turned around and said, actually, no, we don't want the H20.

之後，中國反過來表示，其實我們不想要H20。

And since then, you've publicly declared that you have no China revenue anymore.

從那以後，你們公開表示不再有來自中國的營收。

What was your response to Larnik's comments there?

So first of all, it's really important to understand what AI is.

所以首先，瞭解什麼是AI真的很重要。

What AI is, is a full stack of reinvention of every single layer of the technology stack.

AI其實是科技堆疊中每一層的全面性重塑。

That's the reason why we're talking about energy all the time.

這就是我們為什麼一直談論能源的原因。

When President Trump came into office and he was pro energy growth, if it wasn't because of his posture on energy growth.

當川普總統上任時，他支持能源成長，如果不是因為他在能源成長上的立場。

The United States would have been in a very bad way.

We would have had no energy, no incremental energy necessary to grow this incredibly important industry.

我們將沒有能源，沒有成長這個極其重要的產業所需的增量能源。

So his growth posture, his energy growth policy is incredibly helpful.

所以他的成長立場，他的能源成長政策非常有幫助。

But we have to win every other layer as well.

但我們也必須在其他所有層面獲勝。

字幕列表影片播放

Nvidia 黃仁勳談論AI保麗龍、川普、以及與中國的軍備競賽！ (Nvidia's Jensen Huang on an AI Bubble, Trump, and the Arms Race with China)

incredibly

incredible

engage

access