They overtook Microsoft, Apple, Samsung, Alphabet.

它們超越了微軟、蘋果、三星和 Alphabet。

It's hard to stress just how big of a deal this is.

很難強調這有多麼重要。

I mean, Pokemon, you know, the highest-grossing media franchise in the world?

我的意思是，《口袋妖怪》，你知道嗎，世界上收入最高的媒體系列？

This could buy that company 36 times.

這可以買下該公司 36 倍的股份。

And we know, NVIDIA.

我們知道，輝達。

You might even be watching this video using one of their graphics cards in your PC.

您可能正在使用他們的顯卡觀看本視頻。

But we're talking a company that overtook Apple.

但我們說的是一家超越蘋果的公司。

You know, the guys who manage to sell luxury smartphones to like a third of the world's population.

你知道的，就是那些能把豪華智能手機賣給全球三分之一人口的人。

You don't beat that just with some high-end PC graphics cards.

僅靠一些高端 PC 顯卡是無法做到這一點的。

So what on earth have NVIDIA done?

英偉達究竟做了什麼？

To A, get themselves to a position where they could even dream of overtaking all the other big tech companies.

A，讓自己達到甚至可以超越其他所有大型科技公司的夢想。

And B, actually overtake them.

B, 實際超越他們。

And one thing that I need to make super clear, which makes this all the more surprising, is that NVIDIA only started in 1993.

我需要特別說明的一點是，NVIDIA 公司 1993 年才成立，這讓人更加驚訝。

Which is recent enough that the guy who is now CEO, Jensen Huang, he was the same guy who founded the company along with two other guys.

最近，現任首席執行官黃仁勳（Jensen Huang）與另外兩個人一起創立了這家公司。

Jensen used to work as a waiter's assistant in a local Denny's diner when he was younger.

詹森年輕時曾在當地一家丹尼餐廳當過服務員助理。

He fell in love with computer science and so eventually went from there to getting his dream job working with AMD, one of the leaders at the time as a microprocessor designer.

他愛上了計算機科學，並最終在 AMD 公司找到了他夢寐以求的工作，該公司是當時領先的微處理器設計公司之一。

Until one day when he decided, I got an idea.

直到有一天他決定了，我才有了主意。

Jensen basically realized the key thing that was holding PCs back from unparalleled performance, which I'm getting to.

詹森基本上意識到了阻礙個人電腦實現無與倫比的性能的關鍵因素，這一點我馬上就會講到。

So he booked a table all the way back where it began, at a Denny's diner with these two other guys.

於是，他和另外兩個人在丹尼餐廳訂了一張桌子，回到了開始的地方。

And that is famously where they hatched the plan to form NVIDIA.

他們就是在那裡制定了成立英偉達公司的計劃。

There's literally a plaque on the wall where that meeting happened.

在那次會議召開的地方，牆上掛著一塊牌匾。

But NVIDIA didn't have its name at this point.

但英偉達此時還沒有它的名字。

Do you know where that came from?

你知道這是從哪裡來的嗎？

Well, most of the guy's files documenting their plans about this future company had been pretty uncreatively named, Next Version.

他的大部分文件都記錄了他們對這家未來公司的計劃，這些文件被命名為 "下一個版本"，非常沒有創意。

So when they decided that we're actually doing this, they looked up words that included NV in that order, and they found the Latin word NVIDIA, which is the name of an ancient Roman goddess.

是以，當他們決定真的要這麼做時，他們按順序查找了包括 NV 在內的單詞，並找到了拉丁文 NVIDIA，這是古羅馬女神的名字。

NVIDIA was the personification of envy, a pretty unpleasant lady with a poisoned tongue and green eyes.

英偉達是嫉妒的化身，是一個長著毒舌和綠眼睛的令人討厭的女人。

Thus the name NVIDIA was born.

於是，NVIDIA（英偉達）這個名字就誕生了。

So the brand name basically means Next Version Envy.

是以，該品牌名稱的基本含義是 "下一個版本的羨慕"（Next Version Envy）。

Kind of saying, you're going to be so jealous of our next chip.

有點像在說，你一定會非常嫉妒我們的下一個芯片。

And now that you know where the green eye logo comes from, you too will no longer be able to see it without also seeing this haunting image alongside it.

現在，你知道了綠眼標誌的來源，你也將不再能夠在看到它的同時不看到這個令人不安的影像了。

So you're welcome.

不客氣。

And now that you know, you will see this origin all across their branding too.

現在你知道了，你也會在他們的品牌中看到這個起源。

Like the slogan for the GeForce 8 series graphics cards was Green with Envy.

就像 GeForce 8 系列顯卡的口號 "羨慕的綠色 "一樣。

And the company's highest end GPUs for a long time, they were called Titans.

在很長一段時間裡，該公司最高端的 GPU 被稱為 Titans。

And this NVIDIA was believed by some ancient Greeks to be the daughter of a Titan.

一些古希臘人認為，英偉達是泰坦的女兒。

So it all links together.

是以，這一切都聯繫在一起。

But what actually made NVIDIA different?

但究竟是什麼讓英偉達與眾不同呢？

What was this big idea that they were getting together for?

他們聚在一起是為了什麼大主意？

Well, in 1999, NVIDIA released the GeForce 256, the first chip ever to be marketed as a graphics processing unit or GPU.

1999 年，NVIDIA 發佈了 GeForce 256，這是有史以來第一款作為圖形處理單元或 GPU 上市的芯片。

A piece of hardware that could sit alongside your CPU, which at this time could only process one calculation at a time and take on the graphics rendering load while you're gaming using far more efficient parallel processing.

這種硬件可以與 CPU 放在一起，CPU 一次只能處理一個計算，而圖形渲染負載可以在玩遊戲時使用更高效的並行處理。

It was designed to divide big tasks that would previously be done very slowly into smaller parts and then distribute those parts amongst many, many small processor cores, which in turn freed up people's actual CPUs to focus on things like in-game physics.

它的設計目的是將以前需要非常緩慢才能完成的大型任務劃分為較小的部分，然後將這些部分分配給許許多多的小型處理器內核，這反過來又解放了人們的實際 CPU，使其能夠專注於遊戲中的物理等工作。

The concept of the GPU quickly caught on in both PC and console gaming, which pretty much revolutionized gaming and very quickly turned NVIDIA into a household name.

GPU 的概念很快在 PC 和遊戲機上流行開來，幾乎徹底改變了遊戲行業，並迅速讓 NVIDIA 成為家喻戶曉的品牌。

And NVIDIA has absolutely killed it with GPUs ever since then.

從那時起，英偉達（NVIDIA）公司在 GPU 方面就大放異彩。

Fun fact, actually, this is the graphics card I installed on the first ever PC I built.

事實上，有趣的是，這是我在第一臺個人電腦上安裝的顯卡。

Every couple of years, they've come on stage, they've announced a new generation of graphics cards, and they've been so successful at making each one a leap above the last that they're now outselling their traditional rival, AMD, you know, the company who Jensen originally got his dream job working at, to a ridiculous extent.

每隔幾年，他們就會在舞臺上發佈新一代顯卡，而且他們非常成功地使每一代顯卡都比上一代顯卡有了飛躍，以至於他們現在的銷量已經超過了他們的傳統競爭對手 AMD，你知道，詹森最初就是在這家公司找到了他夢寐以求的工作。

With NVIDIA's first quarter revenue this year at $26 billion compared to AMD's, well, measly 5.5 billion.

英偉達今年第一季度的收入為 260 億美元，而 AMD 只有 55 億美元。

But they didn't get as big as they are today just by coasting on GPU sales.

但它們並不是靠 GPU 的銷量才有今天的規模。

Here's the rocket fuel that made their growth just explode.

這就是讓他們實現爆炸式增長的火箭燃料。

NVIDIA made a huge bet on AI.

英偉達在人工智能領域下了大賭注。

In 2006, NVIDIA released CUDA, Compute Unified Device Architecture, which they could have done a better job naming for something that's actually a bit of a game changer.

2006 年，英偉達™（NVIDIA®）發佈了 CUDA，即 "計算統一設備架構"（Compute Unified Device Architecture）。

It was a software platform that allowed the developers of apps to properly utilize the full power of NVIDIA GPUs, even for non-graphics tasks.

這是一個軟件平臺，允許應用程序開發人員適當利用英偉達™（NVIDIA®）GPU 的全部功能，甚至用於非圖形任務。

Developers could take tasks that would usually be performed one by one on the CPU cores, and instead, just like they've been able to do with games on NVIDIA GPUs, split those tasks up and perform parallel processing using the thousands of GPU cores they now had access to.

開發人員可以將通常在 CPU 內核上逐個執行的任務，像在英偉達™（NVIDIA®）GPU 上開發遊戲一樣，將這些任務拆分開來，利用他們現在可以使用的數千個 GPU 內核進行並行處理。

This was a big moment for apps.

這是應用程序的重要時刻。

It made NVIDIA's graphics cards the tool for much more than just gaming.

這使得英偉達™（NVIDIA®）顯卡不僅僅是遊戲的工具。

They became the gold standard for video editors, for financial modelers, cryptocurrency miners, anyone who wanted to do anything with a computer, and most importantly, for any company wanting to build and train a big, powerful AI model.

它們成為視頻剪輯師、金融建模師、加密貨幣礦工、任何想用計算機做任何事情的人的黃金標準，最重要的是，對於任何想建立和訓練一個大型、強大的人工智能模型的公司來說都是如此。

How do you do that?

你是怎麼做到的？

Not on CPUs, which are doing things one step at a time, but instead by harnessing the full power of a GPU.

這不是靠 CPU 一步步完成的，而是通過充分利用 GPU 的強大功能實現的。

And which GPU are you most likely gonna pick?

你最有可能選擇哪種 GPU？

NVIDIA's.

英偉達的

NVIDIA being the first company to unlock the full computational power of the GPU for the average developer meant that their CUDA software became the default option.

英偉達™（NVIDIA®）是首家為普通開發人員釋放 GPU 全部計算能力的公司，這意味著他們的 CUDA 軟件成為默認選項。

Developers got comfy with it.

開發人員對此已經習以為常。

It improved very quickly.

情況很快就好轉了。

And because this CUDA programming environment was built specifically around NVIDIA's graphics chips, it was super efficient.

由於這種 CUDA 編程環境是專門圍繞英偉達™（NVIDIA®）圖形芯片構建的，是以效率極高。

It's very similar to how efficient iPhones are because every thread of that iOS software is designed around every fragment of the iPhone hardware.

這與 iPhone 的高效率非常相似，因為 iOS 軟件的每個線程都是圍繞 iPhone 硬件的每個片段設計的。

And with this, by being very forward-thinking, by working with other tech companies on AI for literally the last two decades, NVIDIA became embedded into the very fabric of artificial intelligence.

在過去的二十年裡，英偉達一直以前瞻性的思維，與其他科技公司在人工智能領域開展合作，從而融入了人工智能領域。

Whether it's researchers wanting to run complex simulations to figure out what proteins could potentially cure major diseases, or Tesla wanting to train algorithms for their self-driving cars, NVIDIA became the crutch to lean on.

無論是研究人員希望運行復雜的模擬來找出哪些蛋白質有可能治癒重大疾病，還是特斯拉希望為其自動駕駛汽車訓練算法，英偉達都成了他們的依靠。

Literally right now, Tesla's nearing completion on a supercomputer cluster at their Texas Gigafactory, which is supposed to improve Tesla's full self-driving using 50,000 NVIDIA GPUs.

目前，特斯拉德克薩斯州 Gigafactory 工廠的超級計算機集群已接近完工，該集群將使用 50,000 個英偉達™（NVIDIA®）GPU 來改進特斯拉的全自動駕駛技術。

And a sub to the channel would be GPU-tiful.

如果能在頻道中加入一個子頻道，那就太美妙了。

That also could have been better.

這也可以做得更好。

And so when AI development exploded over the last couple of years, with every single person and their dog now having access to tools like ChatGPT, Gemini, and more, what's running all of that?

過去幾年，人工智能的發展呈現爆炸式增長，每個人和他們的狗現在都可以使用 ChatGPT、Gemini 等工具，那麼是什麼在運行這一切呢？

When we ask for an image of a robotic flamingo doing roly-polies on the moon, what's processing that AI request?

當我們要求獲取機器人火烈鳥在月球上嬉戲的影像時，是什麼在處理人工智能的請求？

Yes, NVIDIA.

是的，英偉達。

But another crazy bit of context, NVIDIA's tech is also what powers Amazon Web Services, which is the world's most broadly adopted web services provider.

但另一個瘋狂的背景是，英偉達™（NVIDIA®）的技術也是亞馬遜網絡服務（Amazon Web Services）的動力，而亞馬遜網絡服務是世界上應用最廣泛的網絡服務提供商。

They're the company who builds the infrastructure for basically 30% of the cloud-based internet.

該公司為 30% 的雲端互聯網提供基礎設施。

So the chances are pretty high that anytime you visit a website, there is NVIDIA tech somewhere along the pipeline that's bringing it to you.

是以，無論您何時訪問一個網站，都很有可能是 NVIDIA 技術在為您提供服務。

So that explains the first part of this puzzle, how NVIDIA became the sole arms dealer of the chips required for the now massive AI industry arms race.

這就解釋了謎題的第一部分，英偉達是如何成為目前大規模人工智能行業軍備競賽所需芯片的唯一軍火商的。

But how did they go one step further to actually overtake every single other tech company in June?

但是，他們是如何更進一步，在 6 月份超越其他所有科技公司的呢？

Well, they decided to do a 10-to-1 stock split.

他們決定進行 10 比 1 的股票拆分。

And to find out what that is, I can just ask the ARIA AI baked into the Opera browser.

要想知道那是什麼，我可以直接問 Opera 瀏覽器中的 ARIA 人工智能。

Explain to me a 10-to-1 stock split in two sentences with an example.

請用兩句話解釋 10 比 1 的股票分割，並舉例說明。

Right, so a 10-to-1 stock split means that each existing share of a stock is divided into 10 new shares.

沒錯，10 比 1 的股票分割意味著每股現有股票被抽成 10 股新股票。

And then it gives an example to say that if you owned 100 shares of a company and they did a 10-to-1 stock split, you would now have 1,000 shares in your possession.

然後舉例說明，如果你擁有 100 股公司股票，而他們進行了 10 比 1 的股票分割，那麼你現在就擁有 1000 股股票。

But the total value of your investment is the same because it's just that now each individual share is worth a 10th of what it used to be.

但你的投資總值是一樣的，因為現在每股股票的價值只有以前的十分之一。

Why does that matter?

這有什麼關係？

Well, it means that for new investors, NVIDIA shares have just gotten 10 times more affordable.

這意味著，對於新投資者來說，英偉達股票的性價比提高了 10 倍。

So it makes investing in this company much more accessible at a time where the entire world is looking for some way to invest in AI.

是以，在全世界都在尋找投資人工智能的途徑時，投資這家公司就變得更容易了。

It became a no-brainer.

這變得不費吹灰之力。

It was explosive, as for now and the future.

無論是現在還是將來，這都是爆炸性的。

While NVIDIA has been hit with a pretty severe market correction, so they're back down into third most valuable company.

英偉達（NVIDIA）則遭遇了相當嚴重的市場調整，是以又跌回了最有價值公司的第三位。

And while there are some question marks around AI, like is it being overvalued in the same way that the dot-com bubble was?

雖然人工智能還存在一些問題，比如它是否像互聯網保麗龍那樣被高估了？

And can the world sustain it?

這個世界能承受得了嗎？

Because, you know, there was a recent study that showed that at the current rate of growth, our AI needs alone will consume the amount of power of an entire country the size of the Netherlands.

因為，你知道，最近有一項研究表明，按照目前的增長速度，僅我們對人工智能的需求就將消耗掉整個國家（面積相當於荷蘭）的電力。

But even then, now that the AI cat is out the bag, it doesn't look like AI is going anywhere.

但即便如此，現在人工智能的大勢已去，人工智能看起來也不會有什麼大的發展。

And even if it does, there's no changing that the world is going digital.

即便如此，世界正在走向數字化也是不爭的事實。

Movies are now being made using game engine graphics.

現在的電影都是用遊戲引擎圖形制作的。

Surgeons are now honing their craft using digital twins of patients before operating on the real thing.

現在，外科醫生在給病人做手術之前，都會先使用病人的數字雙胞胎來磨練自己的技術。

The metaverse, if that ever comes to fruition, it's all being powered by NVIDIA GPUs.

元宇宙（metaverse），如果真的實現了，它將全部由英偉達™（NVIDIA®）GPU 驅動。

So that's how NVIDIA got to the crazy position that they're currently in.

是以，英偉達才會走到今天這個瘋狂的境地。

And by the way, ARIA isn't even the most fun thing about the Opera browser anymore.

順便說一句，ARIA 甚至不再是 Opera 瀏覽器最有趣的地方。

So you've seen how tab islands can help you group together all of your tasks into buckets so they're organized.

是以，你已經看到標籤島如何幫助你將所有任務歸類到桶中，使它們井井有條。

But now you can hover over any tab and add a tab emoji too.

但現在，你可以將鼠標懸停在任何標籤頁上，並添加一個標籤頁表情符號。

Why?

為什麼？

Well, aside from a delightful touch of customization, I've had two benefits from this.

除了令人愉悅的定製功能外，我還從中得到了兩個好處。

One, that every time I have two similar tabs open, this allows me to differentiate them.

其一，每當我打開兩個類似的標籤頁時，這可以讓我區分它們。

Instead of persistently gaslighting myself into opening the wrong one every time.

而不是一味地給自己施壓，讓自己每次都開錯。

But also when you're browsing, you tend to have one or two key tabs.

但在瀏覽時，你往往會有一兩個關鍵標籤。

Like for me, it's the document I'm compiling all of my research into.

就像對我來說，這是我整理所有研究成果的文件。

And tab emojis mean that even if you've got like 40 things open at once, I always know to return to my cooking emoji.

而標籤表情符號意味著，即使你同時打開了 40 個東西，我也總知道要回到我的烹飪表情符號。

So give it a go using my link in the description.

請使用說明中的鏈接試試看。