저자: 미상 · 외국인
요약 보기
마이클 버리가 AI 칩 스타트업 Etched를 언급하며 엔비디아 경쟁 위협을 지적했다는 뉴스를 전달. 진행자는 Etched가 약 7억 달러 조달, 약 210억 달러 가치로 주목받지만 엔비디아의 CUDA 생태계·데이터센터 규모·매출 성장(분기 매출 약 816억 달러, 전년比 85%)을 고려하면 단기 위협은 제한적이라고 평가. 버리의 AI 인프라 과잉투자 우려도 소개했으나 Steve Eisman의 반론을 지지. 엔비디아 주가 방향에 대한 명시적 1~3개월 예측은 없고, 고객 이탈·마진 압박·AI 지출 둔화를 지켜보겠다는 입장.
다듬은 글
Something very interesting just happened with Nvidia . And I think anyone who holds shares in Nvidia or any major AI company should understand that. Michael Bury, the investor known for playing against the housing market before the 2008 crash, as we all know, is now pointing to what he thinks could be a serious competitor to Nvidia, a company valued at $5 trillion.
And this time he's thinking about a company called Etched. I'm not saying Nvidia is suddenly out of business or about to go bankrupt. No, this is absolute nonsense. Pay attention to this. But there is a much more important question here. What I'm really trying to answer is : what happens to Nvidia's huge lead in AI if companies start making chips that are cheaper, faster, and designed specifically for new generations of AI models?
What will happen? Because that's what Michael Bury is looking at right now. This could become a very important battle in the entire AI ecosystem if this company becomes a rival to Nvidia. This is what Bury actually said . First, let's understand his words. Bury singled out Etched, an AI chip startup that suddenly became extremely interesting.
According to Business Insider, Etched recently raised about $700 million, valuing it at around $21 billion . This is very impressive. But money is not the most important thing here. Listen carefully. The most important thing is that instead of creating GPUs that can perform different types of calculations, they act differently.
Etched creates chips specifically for transformer- based AI models . According to the report, Michael Bury cited information from an anonymous insider indicating that Etched chips could provide significantly better performance at a lower price for certain workloads. They compare it to Nvidia. Let's leave it at that, because it doesn't mean that anyone has proven they can destroy Nvidia.
This means that someone is attacking Nvidia from a completely different angle. This is what makes the situation so interesting in the context of competition with Nvidia. And why is this so? Because Nvidia's real advantage is huge. Because Nvidia's advantage has never been just in the market. We make fast chips.
Here's what Nvidia says. If everything were that simple, Nvidia would have been replaced long ago. Nvidia has been building an entire ecosystem around its GPUs for years now. You have CUDA, you have developers already working with Nvidia, and there are a lot of them. You have huge data centers designed around Nvidia hardware at scale.
You have a network infrastructure of enormous scale . You have a software of enormous scale . You have companies spending billions of dollars literally building AI infrastructure around Nvidia ecosystems. So this is a huge advantage for Nvidia. So Etched doesn't just need to release fast chips to compete with Nvidia.
They need to convince customers. This is the most important part . They'll ask, " Why should I move some of my workload away from Nvidia? " So it's much more complicated. But here's where things get more dangerous. Etched, a new company, doesn't necessarily have to replace Nvidia overnight. No, it shouldn't.
It just needs to prove that Nvidia shouldn't control almost everything in the AI market. So that's one of the dangerous parts. Now, why do specialized chips matter so much ? Think about it this way. Imagine Nvidia giving you an extremely powerful machine capable of performing many different tasks at once.
Now another company comes along and says: I'm not going to surpass that machine in everything. I will build one machine that performs one extremely important task incredibly efficiently and smoothly. This is a threat to Nvidia because the AI industry is becoming so huge. Companies spend extraordinary amounts of money on computing systems.
Microsoft, Amazon, Alphabet (Google), and Meta are just a few of the companies investing enormously in AI infrastructure, with aggregate spending approaching hundreds of billions of dollars in just a few quarters, and at such scale, efficiency becomes extremely important . If someone can really reduce the cost of running certain AI models while still providing competitive performance, companies will listen to them.
Do you understand what I mean? Not because they hate Nvidia. It's not because of that. Because saving even 10% on huge infrastructure costs can mean billions of dollars in savings. This is where the competition gets so serious for Nvidia, if this comes to fruition soon. Now, Michael Burry's broader argument regarding Nvidia is a bit different.
But understand something else. Michael Burry did n't suddenly become pessimistic about Nvidia. No, not because he discovered Etched yesterday. Just yesterday. This is part of a much larger argument he makes against AI trading. Michael Burry has already revealed his bearish stance on Nvidia and other AI companies, repeatedly arguing that spending on AI infrastructure could become very, very excessive over time.
He has many complaints against many AI companies. His concern is mainly this. Everyone is spending huge amounts of money because no one wants to be left behind in the field of AI. This is his simple argument. Nvidia sells chips. Data centers are being built. More and more funding is pouring into this AI system.
More GPUs are ordered every day, every week, every month , and this cycle continues. But ultimately, according to Michael Burry's arguments, someone has to prove that all these costs provide sufficient economic returns. This is what Wall Street wants. Now add competition to that. If Nvidia customers eventually have more choice, what will they do?
Nvidia could face pressure in areas where it currently has extraordinary pricing power. This doesn't mean Nvidia will go bankrupt, but Wall Street doesn't need the company to collapse for the stock to run into trouble. Wall Street doesn't need this. Wall Street just needs things a little cheaper. And this is where the score matters a lot for Nvidia.
This is extremely important because stock prices are formed based on more than just whether a company makes money. They are not valued that way. They are valued based on what the market expects from the company in the future. Looking at every report, we see that Nvidia has shown extraordinary growth over the past few years.
Surprisingly. Revenue for the first fiscal quarter reportedly reached about 81.6 billion, up 85% year-on-year, while data center revenue grew even faster. So, these are huge numbers for Wall Street. This is no joke. So, Michael Burry is not looking at a weak company that is losing money. He is actually pitted against one of the strongest businesses in the entire AI boom.
This is Nvidia. That's what makes this discussion so interesting. The bearish argument is not that Nvidia has no business. That's not the point. Of course he is. Nvidia has a lot of business, a lot of customers, a lot of reliable stuff. The bearish argument is this. How long can Nvidia's dominance last as competition grows stronger?
Because if the market expects near-perfect execution, even a small change in expectations can dramatically turn the stock in the other direction. Did you understand my point? So not everything has to be perfect. But here's the biggest problem with Burry's argument. What I see now is that we have to be objective.
Just because Michael Burry is worried doesn't automatically mean he'll be right. No , it doesn't mean that. Now there is a serious counterargument. Steve Eisman, another investor associated with the infamous housing market prediction, has a very different view. Let me explain. His main argument is that it's hard to justify a bearish forecast while Nvidia continues to show extraordinary revenue growth.
Yes? And I think I support this argument 100%. And I think that's how we should look at this particular case. Forget about personalities. Forget that Michael Burry predicted 2008 or any other year. No, forget about it. Look at the numbers. This is what Wall Street is looking at. If Nvidia continues to massively grow revenue and demand remains extremely strong, if customers continue to buy the next generation of Nvidia systems, and if Cuda literally keeps customers in the Nvidia ecosystem, then a technically impressive startup won't change Nvidia's entire business overnight.
This is impossible. Listen, Etched still has a lot to accomplish. They need to produce on a huge scale . They have to supply these products. They have to satisfy customers and need reliable customers. They really need reliability. They need software support on a massive scale. So it's a completely different challenge than just building an impressive chip.
You got the point . They can't compete with Nvidia just like that. No. So here's what I'll actually be watching for. Not by headlines like " Nvidia killer is near". I won't watch it. I'm not interested in that. I literally won't pay attention to it. I need evidence. If Etched or another competitor gets a large order, say from a hyperscaler, that's important.
And if companies start to move significant AI workloads from Nvidia, that's important to me; This is extremely important. And if Nvidia's margins start to come under pressure because customers suddenly have alternatives, that's very important to me, but that's not happening. And if AI infrastructure spending slows and competition increases at the same time, that's when that competition becomes much more serious for Nvidia, because Nvidia can compete with its competitors—and it's doing so now, without feeling any negative impact.
Nvidia could potentially cope with the slowdown in AI spending. This is normal. This is business. But tackling both factors simultaneously would create a completely different environment. And remember, friends, Etched is not Nvidia's only competitor. This is not true. AMD, for example, is one of the biggest competitors promoting its own AI accelerators.
Google has its own TPUs. Amazon has its own silicon. And other companies are also developing their own specialized chips. So, Nvidia has huge competition, but Nvidia stands firmly in its place. So the entire industry wants to get a share of this market. Therefore, Nvidia has to move extremely quickly, and it is acting at such a frantic pace.
Now here's what I think. Finally, Michael Burry discovered something worth noting. Okay, I appreciate that. But the fact that Etched has raised about $700 million and achieved a valuation of about $21 billion suggests that serious money believes that specialized AI chips can become a significant business.
But calling this the end of Nvidia would be too premature. So, the real big story is this. For years, the market has viewed Nvidia's dominance as one of the strongest parts of the entire AI investment story. We are now entering the next phase of the AI ecosystem. This is competition. And if these specialized chips really start to deliver significantly better economics for specific AI workloads, Nvidia may have to fight harder for every dollar of AI spending.
So that's something I'm watching very closely. It's not about whether Nvidia will disappear because of this competition. It's not about whether one startup will suddenly come along and replace Nvidia. This will not happen. Not even in a year, and not even in two years. I'm watching to see if Nvidia's incredible pricing power and market position starts to change.
That's what I'm watching. Because if this happens at the same time as AI spending slows, then only then will Michael Burry's warning become much more serious. But that wo n't happen soon. But until we actually see customers switching to other products, margins weakening, or Nvidia's growth slowing significantly, this remains a risk worth watching.
this does not prove that Nvidia has lost its foothold in the AI ecosystem. That's the difference between reacting to a headline and actually understanding the trade. That's a huge difference, guys. I will put the link in the description because we trade every day at the market opening. Come and try it, you might literally make money.
Thank you . See you in the next video, don't forget to subscribe and stay up to date.
Comments 0