JEREMY SIEGEL ON NVIDIA'S $500B DEAL: What It Means for NVDA

요약 보기

시장 변동성 확대와 주요 기술주(Magnificent Seven) 상대 약세를 짚고, 다음 주 연준 회의와 유가 급등(WTI 70→90)이 매파적 우려를 키운다고 봄. AI 에이전트 보안 이슈는 산업 전반의 보안 재점검을 촉발하는 촉매로 보며, 엔비디아의 5,000억달러 AI 인프라 금융 구상은 자산담보 금융 가능성을 보여주지만 실제 집행과 수익성 검증은 남았다는 입장.

다듬은 글

※ 자동 자막 원문(다듬기 전)

01. Tree's chief economist, Jeremy Siegel. Professor Siegel, thank you very much for joining us. As I outlined the market movement and we hinted at what will happen next week with the Fed and interest rates, how much of this market is currently dependent on just a few key sectors and, of course, future interest rate policy.

02. Yes, from both. I mean, some of the fluctuations were just crazy. I think Wednesday was the biggest drop in the "Big Seven" against the S&P in 4 years. Uh, and you know, some analysts are already saying, should we even call them the "Magnificent Seven"? Uh, you know what? Risky competition, uh, margin erosion, uh, and on top of that, as you rightly pointed out, next week is very important.

03. What does Kevin Warsh think about this rise in oil prices? You know, two weeks ago WTI was worth 70, now it's 90. And you, you, you were right , uh, there's a lot of hope, but hope dies last, and we've seen hopes shattered in the past. What does Warsh think about this? I mean, is he willing to see through it or not?

04. And by the way, I wouldn't be surprised if I saw "hawkish" dissent at the meeting on Wednesday. In years past, Professor, we have heard, and perhaps felt some sarcasm about, one word that came out of the Fed and Fed Jay Powell at one time, and that was the word " transient." We don't use it very often these days, but there's still a sense that the transitory or non- transitory nature of inflation will be a focus for the Fed for investors.

05. Is this " hawkish" bias perhaps due to the need to choose a third path to somehow, uh, understand this? What was that wake-up call for you at that moment, as someone completely immersed in technology? Yes, good morning, Nicole . It's always nice to be with you. Thank you for inviting me. I think we are now in this era of zero trust in artificial intelligence agents.

06. And I say this because for years we've been saying: technology is only as valuable as companies actually use it. And the questions used to be : How smart is your model? Well, Wall Street is now saying : where is the return on investment? Where is the money? Because we are seeing AI move from simply providing answers to taking actions.

07. So, investors should now ask companies like OpenAI, Anthropic, and others. Can you control your agents? Thomas Wolf, co-founder of Hugging Face , called it a wake-up call. I call it a catalyst for a couple of reasons, because what happened in, say, three weeks really accelerated the focus on agent safety.

08. Zero trust is, in essence, not just telling an agent what they shouldn't do. Make sure you build the system so that it cannot do anything it is not authorized to do. The next race in AI will be about control. So if we move to zero trust in AI agents, how does that change everything? Because if we see numerous hacks, and they are, how does that change the environment?

09. I mean, are we still…have we even heard calls to stop the development of AI? I mean, you know , this could be one of the extremes, how does this zero trust change now? I think it's, as I said, a catalyst, which I think is a good thing, because the OpenAI disclosure seems to have prompted a lot of other organizations to go back and check their own testing.

10. Anthropic, uh, they reviewed over 140,000 cybersecurity assessment sessions and found three cases involving unauthorized access to external organizations. So, these incidents are more like basic weaknesses. But Thomas Wolf, co-founder of Hugging Face , says it's day one for cybersecurity in the age of agents.

11. If you think about it, one hack is an incident, but when several companies are now revealing and exhibiting similar behavior, it's a signal to the industry, and it should be treated as such. I use the word " catalyst" because I'm a former scientist, but we also see this acceleration. This incident seems to have really accelerated the industry-wide security review of AI agents, and the time has come.

12. I'm thinking about you using the word "exiled," which doesn't necessarily mean malicious, but unknown—for example, that hack happened in early July, if I remember correctly, it was July 8th, and it wasn't detected for another week or 10 days. Uh, that's also problematic. You said that at this point it could actually attract more investment so that AI safety improves.

13. Yes. And the thing is, when it comes to AI and cybersecurity, it's a paradox, because the same capabilities that make AI agents potentially harmful are the same capabilities that make them very powerful cyber defenders. Companies I follow, like Palo Alto Networks Unit 42, call AI—and I think that's a great phrase—a force multiplier for attackers because it makes existing threats faster.

14. It's easier for them to scale . But AI is also becoming an enabler for defenders, Nicole. So, CrowdStrike is talking about the emergence of an agent-based security operation center (SOC). This is a situation where the agent does not simply mark the threat. It can really help to act autonomously, for example, to sort and investigate.

15. There is a startup in the field of AI security called Vector AI that I follow. They are also moving in the same direction of agent-based SOC, where agents can triage cases and prioritize threats so that people can focus on what really matters. So, it's kind of an "agent vs. agent" era. The AI attacks at machine speed, but the AI also defends at machine speed.

16. Is n't that right? That's how I started because I know that's what everyone is talking about. The HACK ETF is interesting. Nvidia's $ 500 billion AI funding play changed everything. Nvidia may have just unveiled something much bigger than just another AI partnership. The company is currently working with some of the biggest names in global finance, including BlackRock, Blackstone, Apollo, KKR, Goldman Sachs, and Brookfield, to help finance over $500 billion in AI infrastructure.

17. And here's the important part. It's not like Nvidia is just spending $500 billion of its own money. The idea is to attract independent long-term capital and create a financial structure around AI infrastructure. Why is this important? Because Jensen Huang is making a much bigger argument than just “we need more GPUs .” He says that computing power itself is becoming an investable infrastructure asset.

18. nergy, land, networks, and equipment. But now the computing power inside this facility can actually generate revenue. And Nvidia believes its AI infrastructure has three characteristics that investors are looking for . She is productive. It generates revenue and can potentially be refocused on other customers.

19. The last part is especially important. If one AI company stops using a particular data center, the infrastructure doesn't necessarily become useless. Other cloud providers, enterprises, or AI companies can potentially use this computing power, and this is what makes asset- backed financing possible .

20. Here's where it gets interesting for Nvidia investors: For years, one of the biggest questions about Nvidia has been how long these AI investments can last, but these announcements potentially add another source of fuel instead of hyperscalers having to fund the entire build directly from their own balance sheet.

21. The capital market can help finance this, and there is a huge amount of capital looking for productive assets. The CEOs at this table are definitely not selling lemonade. We are talking about some of the world's largest asset managers. But there is an important nuance. This does not mean that $500 billion is guaranteed to be used tomorrow.

22. These are partnership and financial initiatives . And not a single check for $500 billion. And investors still need to answer a more important question. Will the calculations provide sufficient returns to justify the investment? Currently, demand remains extremely high. Data centers need more energy, and AI companies need more computing power.

23. Hyperscalers continue to increase capacity, and the Nvidia ecosystem provides infrastructure flexibility that will make assets more useful in the long term. But eventually supply will catch up with demand, and then investors will have to separate the winners from the companies that simply built too much capacity.

24. That's why I think this Nvidia announcement is more than just another optimistic headline. This is the potential beginning of a new financial level in the AI ​​economy. AI infrastructure is no longer treated simply as technology. It is increasingly seen as an infrastructure capable of attracting debt, equity, and institutional capital.