Name: Meta 將部署數百萬顆輝達（Nvidia）處理器！ (Meta Will Deploy Millions of Nvidia Processors)
Uploaded: 2026-02-19T21:33:26.000Z
Duration: 4 min 35 s
Description: 【看影片學英語】數萬部 YouTube 影片，搭配英漢字典即點即查，輕鬆掌握單字發音與用法，長久累積看電影不必再看字幕。

Metta planning to deploy millions of Nvidia processors in the next few years.

Metta 計劃在未來幾年部署數百萬個 Nvidia 處理器。

A deal that could add billions of dollars to Nvidia's balance sheet.

這項交易可能會為 Nvidia 的資產負債表增加數十億美元。

Bloomberg Intelligence Global head of tech research Mandeep Singh joins us now.

Bloomberg Intelligence 全球技術研究主管 Mandeep Singh 現在加入我們。

Mandeep, first of all, I mean, does Nvidia need this?

Mandeep，首先，我是說，Nvidia 需要這個嗎？

It's up pretty sharply off the back of this.

I thought there was already like pretty robust demand for what Nvidia is selling.

我以為 Nvidia 賣的東西已經有相當強勁的需求了。

And look, this is probably matters way of locking in that Nvidia supply.

而且你看，這可能是 Metta 鎖定 Nvidia 供應的方式。

So not only are they using the GPUs, but they are also talking about using Nvidia's C CPUs.

所以他們不僅在使用 GPUs，還在討論使用 Nvidia 的 CPUs。

I mean, traditionally all These companies used x86 architecture and Intel CPUs along with the GPU cluster that Nvidia had.

我覺得，傳統上這些公司都是使用 x86 架構和 Intel 的 CPUs，還有 Nvidia 的 GPU 集群。

But now, you know, with this deal, they're talking about using Nvidia CPUs.

但現在，你知道的，有了這筆交易，他們在討論使用 Nvidia 的 CPUs。

The other thing for Nvidia here is Anthropic has had a very steep rise in the last few months.

對 Nvidia 來說，另一件事是 Anthropic 在過去幾個月裡有非常急劇的上升。

Entropic models aren't trained on Nvidia GPUs.

Anthropic 的模型並不是在 Nvidia GPUs 上訓練的。

它們是在 Google, GPUs, Amazon 上訓練的。

So from that perspective, you have a company that's doing really well on the model side that is not trained on Nvidia GPUs.

所以從這個角度來看，你有一家在模型方面做得非常好的公司，但它並不是在 Nvidia GPUs 上訓練的。

And so from that perspective, I think for both of them it's important that you know, Metta comes up with a model that is really good in terms of being a frontier model like Anthropics that is trained on Nvidia's latest cluster and obviously matter needs to lock in the supply.

是以，從這個角度來看，我認為對他們雙方來說都很重要，你知道的，Metta 必須推出一個在邊緣模型方面表現非常好的模型，就像 Anthropic 的模型一樣，而且是在 Nvidia 最新的集群上訓練的，顯然 Metta 需要鎖定供應。

That's a good point too because also Gemini, I mean of course they're yeah, using Google's.

說得也對，因為 Gemini，我的意思是，當然他們是，你知道的，使用 Google 的。

I mean in all of this, can you will other companies look at the advances of Anthropic, of Gemini and start to rethink their supply chain because of that, that they don't want to be too reliant on Nvidia.

我的意思是，在所有這些情況下，其他公司會不會看到 Anthropic 和 Gemini 的進展，然後因為不想過度依賴 Nvidia 而重新思考他們的供應鏈？

what Nvidia is saying is our latest Blackwell chips will give you 30x more token output than our prior architecture from Hopper.

Nvidia 說的是，他們最新的 Blackwell 晶片，與上一代的 Hopper 架構相比，能提供 30 倍的 token 輸出。

So for the same unit of power you can have 30 times more output.

所以在相同的功耗下，你可以獲得 30 倍的輸出。

And so that throughput aspect of Nvidia chips is always there and that ties into how good the model will be that is trained on the latest chip.

是以，Nvidia 晶片的這種吞吐量方面一直都在，這也關乎到在最新晶片上訓練的模型會有多好。

So really what Nvidia needs is companies like OpenAI, like Metta training their models on Nvidia's chips and having somewhat of an advantage compared to the ones that are not trained on Nvidia.

所以 Nvidia 真正需要的是像 OpenAI、像 Metta 這樣在 Nvidia 晶片上訓練他們模型，並且相較於那些沒有在 Nvidia 上訓練的公司，能獲得一些優勢。

Where do some of these other competitors stand?

I think AMD gets hit even with this tie up between Metta and Nvidia, where Nvidia is now supplying the CPUs.

我認為 AMD 即使在 Metta 和 Nvidia 的合作中也受到影響，因為 Nvidia 現在正在供應 CPUs。

Previously AMD or Intel was the supplier of CPUs to the data centers for Metta.

以前 Metta 的資料中心 CPU 供應商是 AMD 或 Intel。

Now with Nvidia selling CPUs that could have an impact on AMD.

現在 Nvidia 也開始賣 CPU 了，這可能會影響到 AMD。

Are you surprising that a company like intel that has a government backing there haven't been more political forces?

你不覺得像 Intel 這樣有政府撐腰的公司，政治力量沒有更大一點嗎？

Because this is what was talked about, right?

因為這就是大家討論的重點，對吧？

The US takes a stake and pushes companies to start using intel for things like CPUs.

美國政府會入股，然後推動公司開始使用 Intel 的 CPU 等產品。

Are you surprised that we haven't really seen that?

你不意外我們沒看到這種情況嗎？

Well, intel still has the foundry capacity.

嗯，Intel 還是有晶圓代工產能的。

So the difference between an intel and AMD is intel owns its foundries, where it makes its chips.

所以 Intel 和 AMD 的差別在於，Intel 自己擁有晶圓廠來製造晶片。

所以從這個角度來看，你說得對。

I mean intel, it has a local foundry that all these chip makers could use.

我是說 Intel，它有在地晶圓廠，所有這些晶片製造商都可以使用。

But obviously the government role comes into play in terms of pushing these companies to start using Intel Foundry more and more.

但顯然政府的角色在於推動這些公司越來越多地開始使用 Intel Foundry。

So just the one other concern that we have to bring up, as we do with any of these conversations, is the circularity of deals.

所以我們在進行任何這類對話時，都必須提出的另一個擔憂，就是交易的循環性。

The fact that Metta already makes up a nice chunky portion of Nvidia's revenue, presumably that just reinforces that you say it's a good thing for Nvidia.

Metta 已經佔了 Nvidia 營收的很大一部分，這想必也加強了你說對 Nvidia 來說是好事。

But is there also kind of like a dark side to the silver lining?

但這光明的背後，是不是也有點陰暗面？

I mean, look, all these companies have an ecosystem.

我是說，你看，所有這些公司都有一個生態系。

So that circularity also exists in, you know, Google investing in Anthropic and Anthropic using Google's chip.

所以這種循環性也存在於，你知道的，Google 投資 Anthropic，而 Anthropic 使用 Google 的晶片。

Same thing with the Amazon Anthropic partnership.

Amazon 和 Anthropic 的合作也是一樣。

And at the end of the day it is an ecosystem because you're optimizing on a certain chip vendor for your model training and your model inferencing and you can't have a generic architecture.

歸根究柢，這是一個生態系，因為你針對特定的晶片供應商來優化你的模型訓練和模型推論，你無法擁有一個通用的架構。

And that's why, you know, they are very keen to partner with the companies of meta scale.

這也是為什麼，你知道的，他們非常熱衷於和這些大型公司合作。

You want to make sure they optimize on Nvidia chips as opposed to using someone else's.

你希望確保他們在 Nvidia 晶片上進行優化，而不是使用別人的。

你的世界裡永遠都不會有無聊的一天。

Bloomberg Intelligence 的 Mandeep Singh。

You probably wish for a boring day that might be nice.

你可能希望有平淡無奇的一天，那樣也不錯。