저자: 댄아이브스 · 한국인
요약 보기
AI 군비 경쟁이 핵심이며 엔비디아 독주에 균열이 생기고 있으나 여전히 AI 칩 시장을 지배한다고 주장 메타·앤트로픽의 구글 TPU 도입 협상이 엔비디아 독점에 대한 신호이며 TPU는 엔비디아보다 3~4년 뒤처지지만 특정 용도에서는 가성비가 좋다고 평가 칩 수요/공급 비율은 12대1, 엔비디아 주가는 경쟁 뉴스에 일시 하락 AI 수요는 6개월간 30% 가속, 미국 기업 중 AI를 제대로 도입한 곳은 3%에 불과해 현재는 1회 초 종료 단계라고 주장 AI 지출은 주로 기업(B2B) 중심이며 소비자 AI 혁명은 아직 도래하지 않았다고 평가 미국이 30년 만에 처음으로 AI에서 중국을 앞서고 있으며 AI가 반도체 기술 경쟁의 핵심 전장이라고 진단 현재 강세장은 4차 산업혁명의 초기이며 최소 2년 이상 지속될 것으로 전망 애플은 23억 4천만 대의 iOS 기기를 보유한 플랫폼 강점이 있으며 구글 제미나이 파트너십을 통해 주당 75~100달러의 추가 가치 창출 가능, iPhone 17 교체 사이클 기대 메타는 최근 20% 하락했으나 CAPEX 확대는 생존을 위한 필수 투자이며 30억 명 사용자 기반의 수익화 잠재력이 크다고 강조, 강력한 매수 기회로 제시 테슬라는 자동차가 아닌 AI·로봇 기업으로 평가, 완전자율주행과 옵티머스가 핵심이며 규제 완화로 미국 30~35개 도시에서 로보택시 운행 전망, 웨이모는 반올림 오차 수준이라고 평가 결론적으로 AI 붐은 단기 거품이 아닌 장기 산업 혁명이며 빅테크의 막대한 투자는 시장 주도권 확보를 위한 필수 전략이라는 입장
다듬은 글
. I have brought a really interesting topic today. yes . hello. He is a famous analyst at Wedbush Securities. Ben Ives defined the current tech market in just one word. It is the AI arms race. An arms race? Wow, that expression is really intense. Right? So today, we are going to step right into the middle of that war .
Based on the materials sent by our listeners, we will delve deeply into who the companies currently leading the competition are and how their strategies will transform the technology landscape . yes . Actually, everyone is only talking about Nvidia, Nvidia. that's right. I think the biggest point of interest will probably be whether a force can emerge to stop this solo run.
That is exactly it. And I'm also curious about what kind of big picture these giants like Apple, Meta, and Tesla are drawing . Our role is to use this session to clearly summarize the complex landscape of the AI industry so that you, our readers, can fully understand it . great. So, shall we start off with some very exciting news?
You know, there's the meta platform. Yes, Meta. Meta calls it Google's AI chip, the TPU . There are reports that negotiations are underway for billions of dollars—that is, trillions—to purchase this . Uh, I saw that news too. This is n't just simple news. That's right . I don't think this is just at the level of Meta buying chips.
Could I perhaps view this as the signal that breaks the ice on Nvidia's once-impregnable stronghold for the first time ? yes. You can certainly see it that way. This is a really important move. You should view it as having significance beyond a simple parts purchase contract . In what sense? for a long time.
For a really long time, AI chips are synonymous with NBD. This formula has never been broken. That's right. It was just the only option. yes . However, the very fact that a big tech company like Meta is seeking alternatives, and doing so openly, sends a tremendous message to the market . Ah, so it's a kind of declaration .
you're right. I think this is a very symbolic event that, for the first time, demonstrated the potential for Google's TPU to become a substantial competitor capable of directly challenging Nvidia in the market, rather than just an experimental chip used only internally. Come to think of it, until now, everyone just didn't bother because they could n't get an Nvidia GPU, and I've hardly ever heard of any other options being available .
But I'm curious, does Google TPU really possess that level of technical capability? Was it receiving that much recognition in the market? You can view this as the process being proven by this incident. Actually, this isn't a story that just popped up out of nowhere . Google already has Antropics, which is very famous as an AI model developer .
yes . It's the company that made Antropic Cloud, right? yes that's right. We have already signed a contract to supply up to 1 million TP chips to Antropic . Seaport analyst J. Goldberg made a comment regarding this contract, calling it a really strong validation of the TPU. They evaluated it as really powerful validation .
Wow, really strong verification? yes . Ah, so, it seems that the company making the best AI models declared that they would develop Ichibo, and the fact that it is THAAD in large quantities has effectively served as a kind of quality guarantee . That's correct. It means their technological capabilities have been properly proven in the market .
However, if a deal with Meta, who spends money on top of Jordan, is also concluded here. Wow, then the impact must be huge. That's right . Then, TPU will no longer be at the level of mere theory, but will firmly establish itself in the market as a very realistic and feasible alternative to NBD chips .
Geureokjun. Of course, Google is still taking a somewhat cautious stance. Officially, they gave a somewhat diplomatic response, stating that demand is increasing for both their own TPUs and NVIDIA GPUs, and that they are committed to supporting both options . Well, of course they would say that. There is no need to go out of your way to stimulate the NBD.
That's right . All right. Google's rise is really interesting, but realistically speaking, Nvidia still completely dominates the market. How does Dan Ives view NBDI's current status ? That is precisely what makes this situation more three-dimensional and interesting. Denaves still states very emphatically that there is only one chip driving the AI revolution right now, and that it is solely Nvidia .
Firmly? yes . The numbers he presented are particularly shocking. Currently, the supply-to-demand ratio for Nvidia chips has reached a whopping 12 to 1. wait a minute. Once again. You're saying it's 12 to 1? yes . It means that 12 companies are competing for a single chip . and. It means 12 people are standing in line to buy one chip .
It’s not just that it’s popular; it’s practically desperate . That's right . So, Meta kept knocking on Nvidia's door, but when that didn't work, he ran to the house next door, Google. It would be fine to view it this way . That is correct. Dan. That is exactly the same as IBS's analysis. The fundamental reason Meta is knocking on Google's door is that it simply cannot secure the desired amount of chips from Nvidia when it wants them .
Ah, paradoxically, this serves as proof once again of just how formidable Nvidia's market dominance is . That's true. It turns out the reason competitors are rising is simply because the number one spot is doing too well . But I have a question. Even though there was a huge fuss about the lack of new stock, I heard that Nvidia's stock price actually wavered a bit when the follow-up news regarding Meta and Googol came out .
How should I interpret this ? Dan IVC said that he was also surprised when he saw that very spot. Oh, really? yes . Everyone thinks everyone is clamoring for Nvidia chips, but the truth is, Nvidia's stock faced selling pressure due to news of its competitor Google selling chips. That's strange. I think this shows that there is a slight discrepancy between market participants' belief in overwhelming demand and the actual stock price movements reacting to short-term changes in the competitive landscape .
It also means that the market is reacting extremely sensitively to every single piece of AI-related news . All right. Of course, price and supply issues are important, but ultimately, it's performance. Is Google's TPU technically capable of really replacing Nvidia GPUs? That is a very good question.
If you look solely at technical capabilities, IVC gives a rather cold assessment that Google's TPU technology is about 3 to 4 years behind Nvidia. Three or four years? The grid is quite large. yes . It's not small. However, I added a very important clue here. In certain cases, it can be quite good. Ah, is it a specific case?
yes . This is the key point. The point is that you don't need to surpass Nvidia in every aspect of every task . Ah, I understand. So, in general, Nvidia is the best, but for the specific purpose we intend to pursue, Google chips can also be incredibly cost-effective. So that's what it means. that's right.
Consider the situation of the memories that are currently facing setbacks in AI model development or service expansion due to the lack of chips . This means that Google TPUs, which can be reliably supplied in large quantities even at the cost of a slight performance difference, can be a very reasonable and attractive alternative .
okay. Seeing a situation like a war unfolding over a single chip, I can understand, in a way, why the term "AI bubble" is being used. In particular, people like Michael Burry from the movie The Big Short viewed it quite skeptically . yes . There have been such concerns for a while. But Dan Ibus said that he viewed this fever not as a bubble, but rather as merely the beginning .
Far from decreasing, demand is accelerating. yes . Ibus shared his story based on his firsthand experience of staying in Asia for three weeks recently . It was analyzed that the demand for AI has accelerated by as much as 30% in the last six months alone . 30% in 6 months? yes . This is a really incredible speed.
But there is something else that is really important here . It is the percentage of companies that have adopted AI in the overall market. Oh, how much is that? According to his analysis, only 3% of all companies in the U.S. are currently following the proper path to adopting AI technology . Wait a minute.
Only 3%. That means 97% have n't even started yet. you're right. and. I thought it would be at least 20 or 30%, but it is much earlier than I expected . that's right. They even say that in other Asian regions, excluding Europe and China, the figure is close to zero . You mean it's close to 0? yes. What would this mean?
All the enthusiasm, investment, and commotion we are witnessing right now is merely the prelude to a massive AI revolution, and full-scale growth has not even begun yet. and. To borrow De Ives' ratio, it is like the top of the first inning of a nine-inning baseball game has just ended. The top of the first inning just ended.
Wow, it really feels real now. Then where on earth is that enormous AI-related spending, which the media is currently talking about as reaching $2 to $3 trillion, coming from? It doesn't mean that 3% of companies spend all that money. It is very important to clearly identify the entity responsible for the expenditure .
Ives explains that current AI spending is almost entirely centered on the enterprise market, or the B2B market. Ah, it is money spent by companies. yes . The analysis is that the consumer AI revolution has not yet arrived . We are currently at a stage where companies are pouring massive amounts of money into building infrastructure, increasing internal efficiency, and developing new business models .
They also pointed out the geopolitical perspective. There was also talk of the hegemonic rivalry between the U.S. and China. yes . That part was also a very interesting point. He assessed that the United States is clearly ahead of China in the field of AI for the first time in the last 30 years . For the first time in 30 years.
yes . This means that AI is becoming the key battleground in the semiconductor technology competition. [Snicker] In addition to this, it was mentioned that the trend of sovereign nations, such as sovereign wealth funds in the Middle East, actively investing in AI with the fate of their countries on the line is also a very important variable.
To sum it all up, this isn't just a simple technology trend, but a massive change on the scale of an industrial revolution, which is why you expect it to last for a very long time. That is correct. Denaves does not view the current tech bull market simply as a cyclical market trend. They are defining it as the Fourth Industrial Revolution.
The Fourth Industrial Revolution yes . We assess that we have only just entered the third year of the massive construction period spanning eight to ten years . So, I strongly predict that this bull market will continue for at least two years, or perhaps much longer . and. Looking at the macroscopic picture like that, I can somewhat understand why individual companies are investing so desperately.
[Snoring] Now, let's take a closer look at the inner workings of each company . First up is Apple. yes . apple. Ives made a very funny remark that Apple has been watching the AI competition from the stands until now, and it seems like they are finally getting ready to come down to the arena. What card will Apple play?
Apple's biggest and most powerful weapon is something no other company can imitate. It is precisely the overwhelming installed base, that is, the platform. A platform? yes . There are 2.4 billion active iOS devices worldwide. They want 100 iPhones among them. 2.4 billion outline? Wait a minute. To give you an idea of the scale, it means that nearly one in three people worldwide uses Apple devices .
Right? and. It seems like a superhighway for distributing AI technology to consumers has already been perfectly laid out . That is exactly it . No matter how outstanding the AI technology other companies develop, how do we deliver it to users? If this is something that needs to be considered, then Apple does n't need to worry about it.
That's true. And Ibus presented the value of that highway in very specific numbers . Are you saying it can add an additional value of $75 to $100 per share to the stock price ? Wow, 75 to 100 dollars. It’s not just that it will be good, but that it is very specific. Yes, it is very specific. To make that number a reality, the best AI model is ultimately needed, so the prediction that Apple will partner with Google's JAME AI rather than build it themselves is really interesting.
It is a bit surprising that proud Apple is using a competitor's technology. That is exactly Apple's very clever and strategic judgment. Google recently won the factual victory in the search engine lawsuit against the U.S. Department of Justice (DOJ) . Ah, yes. That's right. As a result, analysts suggest that the legal and regulatory uncertainties associated with Apple's partnership with Google have been significantly resolved .
Of course, there is the option of acquiring a promising startup like Perplexity, but the point is that incorporating Google Jaminai, an already proven, top-tier model, is the fastest and most certain way to provide the best AI experience to billions of users . That could be a wise decision. yes . And we have very high expectations that this AI integration could lead to a massive upgrade cycle that we could not have imagined, starting with the iPhone 17 .
The iPhone 17. All right. Next, let's move on to Meta. With stock prices recently dropping as much as 20% from their peak, investors must have been quite anxious. yes . That's right. However, Dan Ives actually recommended that now is a strong buying opportunity for a table founder buy that needs to be hit .
I told you to buy when everyone else is scared and selling, so what on earth are you basing your argument on to speak so strongly ? The cause of the stock price decline is very clear. This is because CEO Mark Zuckerberg announced that he would significantly increase capital expenditures, or CAPEX, to avoid falling behind in the AI arms race.
CAPEX capital expenditure? Yes, yes. This places a huge burden on the company's revenue and cash flow in the short term. Wall Street hates seeing numbers get worse right in front of them the most, so stock prices reacted instinctively . That's right . Because they prioritize current performance over the future .
you're right. However, Ives values that very point highly. I viewed this increase in CAPEX as a wise move that was absolutely necessary for long-term survival and market dominance, even if it reduced short-term pain . A wise move. But Wall Street's perspective is the exact opposite. From an investor's perspective, it might sound a bit frustrating to be told to invest with only the future in mind when you can't make money right away .
That's right . Do you think Zuckerberg is really showing a vision worth enduring that short-term pain ? Dan Evils sees it that way. They even highly praised Zucker's leadership, comparing him to an exhibition CEO who makes decisive decisions in crisis situations . Exhibition CEO. yes . It means Zuckerberg knows more accurately than anyone that if he skimps on investment now, his presence in the market could disappear in a few years .
Ultimately, the core investment logic for Meta is likely the expectation that its massive user base can eventually be monetized through AI . That is correct. It's a massive ecosystem used by 3 billion people combined, including Facebook, Instagram, and the Watch app . 3 billion people? Yes. If we start generating revenue in earnest by integrating new AI-powered advertising models, commerce, and paid features on this platform, the potential will truly be beyond imagination .
The massive investment now is a new spring of seeds for that future. All right. Finally, let's move on to the story of Tesla. Honestly, looking at this year alone, it seems like investors must have been really anxious . Yes, it was very difficult. Everyone is worried about things like the slowdown in EV sales, but you're saying Dan Ives sees value in a completely different place ?
The perspective that it is not a car company. I'll have to look into this a bit more deeply . that's right. He believes that Tesla's true value lies in the much bigger picture of AI and robotics, beyond the automobile manufacturing process. AI and Robotics. yes . In particular, I say with absolute certainty that fully autonomous driving technology and the humanoid robot Optimus will become the most important chapter of Tesla's future growth story .
The most important chapter? yes . The point is that electric vehicles were just the starting point for realizing this massive vision . The outlook for autonomous driving, especially robot taxis, was also quite concrete. Actually, there is also a loss of trust because they have been repeating the same promise that it will happen soon for years .
yes . There is a need for that part. However, Ives assessed very positively the fact that the regulatory issue, which had been the biggest obstacle so far, is finally showing signs of cooling down . Ah, the regulations are being lifted. yes . He predicts that as the regulatory roadmap for building robot cities becomes clearer, we will see robot taxis actually operating in 30 to 35 cities across the United States in the future.
30 to 35 cities. If that is decided and becomes a reality, then so be it. Tesla's corporate value will inevitably be re-evaluated on a completely different level from now. Still, the competitors are formidable. In particular, companies like Google are receiving significant recognition for their technological capabilities.
Ives is very firm about that part . How did you say it? He even went so far as to say that competitors like Waymo would be nothing more than a fifth-order rounding error compared to the massive vision Tesla is envisioning . Rounding up to the 5th? Wow, that's a strong expression. I firmly believe that Tesla's approach to developing technology independently will eventually create an overwhelming technological gap .
After listening to the whole story today, I really feel that there is a massive paradigm shift . Yes, I feel the same way. To summarize, Dan Ibus's analysis is that the current AI boom is not a short-term bubble, but the Fourth Industrial Revolution that will continue for several years to come . yes .
That's the key point. And so, the astronomical investments we see from Big Tech are an essential survival strategy to seize control of the future market. Yes, that is correct. To summarize the key points, it is like this. The market has begun to undergo a tectonic shift as alternatives like Google emerge amidst Nvidia's dominance and supply shortages .
yes . AI technology adoption is still in its very early stages, with only 3% of all companies participating . That's right . Top of the 1st inning. yes . And Apple is preparing for this revolution in its own way through a massive consumer platform, Meta through monetizing its 3 billion user base, and Tesla through the transition beyond cars to AI and robotics .
Understood. Amidst this massive flow, I would like to wrap up by posing one final question for our listeners to reflect upon . yes . great. Today, we mostly finished recommending how existing big tech giants will navigate this competition . that's right. However, as Dan Ivos says, if we are only in the third year of a massive revolution spanning ten years, could n't we ask this question?
What kind of question is it? Is n't there a possibility that a new startup, currently off our radar at all, or a disruptive entity could suddenly emerge and completely shake up the very Big Tech-centric landscape we discussed today? and .
Comments 0