저자: DanIves · 외국인
요약 보기
Dan Ives는 아시아 공급망 체크 데이터를 근거로 엔비디아 실적 추정치가 25-30% 과소평가됐으며 향후 6~8분기 실적이 상향될 것이라 주장. 수요 대 공급 비율이 10:1에서 13~14:1로 확대됐고 엔비디아가 AI 칩의 유일한 공급자라고 강조. 마이크로소프트는 Azure 기반 엔터프라이즈 AI 수주 확대로 월스트리트가 반영하지 않은 매출 10~15% 추가 여지가 있다고 봄. 이후 자막 후반부는 AI 인프라 자금조달 우려(CoreWeave, Oracle 부채 부담)와 사모 대출·칩 담보 가치 관점을 다루며 엔비디아·브로드컴·마이크론은 자금조달 의존도가 낮아 차별화된다고 설명.
다듬은 글
Managing Director Dan Ives. Nice to see you, friend. Thank you for coming back. Nice to see you. What do you think is behind this recent stock rally? I think the most important thing is that, judging by all the checks coming out of Asia, investors are starting to understand that the figures are probably understated by 25-30% for next year.
I mean, we're seeing a huge acceleration with all these enterprise AI use cases, and Nvidia is the only chip in the world that's powering the AI revolution. This is what we are observing now. Undervalued by 20-30%. Really? That's a lot. How could we be so wrong? Yes. Demand for chips has skyrocketed, and if you look back at the last few quarters, Nvidia has been treated somewhat indifferently.
But if you look at what's happening in Asia, according to all our data, things are accelerating. The demand-to-supply ratio has changed from 10:1 to 13:1, 14:1, because now physical AI and other enterprise use cases—all of that is expanding. So it's not just about hyperscalers. It all comes down to the fact that there is only one chip in the world that delivers this, and that's Nvidia.
And I think investors were somewhat cautious last quarter, and now it's becoming quite obvious that the numbers are significantly understated not only for this quarter, but probably for the next six to eight quarters. Perhaps investors are betting on this, pushing the NASDAQ to a record high again. I want you to listen to Josh Brown's commentary from today's program, about mega-cap companies, why they're growing, and where they might be headed next.
But the truth is that if Apple is in a product cycle and reaching peaks, and Nvidia, after consolidating for God knows how long, starts to renew highs, and at the same time, profit forecasts for these two companies increase. Microsoft is in a state of stagnation. Meta has regained its former form. They regained control of the narrative.
They get multiplier expansion in addition to a new earnings growth story. These are the stocks that really matter to how the S&P ends the year. I assume you agree with that . I don't think this is a big surprise , given the size of the market capitalization of these companies. But regarding the idea with Microsoft, which, by the way, received a rating upgrade from Melius today, the target price is 665.
You know, over the past month, apart from Meta, Microsoft is the best performer , although it is slightly ahead of its competitors among mega-capitalizations. Why? By itself . That's because, that's because it concerns businesses. Because when you look at Azure and where these use cases are going, ultimately it's Redmond and Satya Nadella who are winning so many of these deals.
And I think that goes back to the last few quarters; Look, Microsoft was one of those that, you know, everyone reacted very negatively to, but as you can see now, as the use cases expand, your numbers could be pointing to an additional 10-15% of revenue that Wall Street isn't accounting for. And I think all of this ties into the Nvidia theme and what we're seeing in hyperscale as they accelerate over the next few quarters.
Big tech companies are showing their strength again, and when you add Meta and others, I think it reinforces the idea that investors who are betting against these companies, whether it's Apple or Meta, have a user base that is a golden goose, and that's what these companies have. 87 versus 85 expected, and if you look at the revenue of 21.87 billion, it turned out to be lower than the almost 22.3 billion they were counting on.
So, we have a miss in the total income indicator. Net interest income—a slight deviation: 12.52 billion versus 12.63 billion expected. And then they look at the net interest income forecast, about $50 billion. This was about within normal limits, although the current quarter is expected to be slightly smaller than anticipated.
Income from investment banking activities amounted to $611 million. So they don't rely on trading as much as big banks or maybe investment banks. But total revenue from markets was $1.84 billion. They gave a fairly optimistic forecast regarding consumer sentiment and the state of the economy in general.
So, if you look closely, the prospects are better. Shares fell about 10%. Slightly down this morning. Pre-market indicators have retreated from the recent highs we have seen. So, yes, they gave a good forecast for consumers. They rely more on the credit side of the business. They also had some loan loss provisions that they had to increase a little bit , but overall it was a mixed quarter.
It's no surprise that we see a little weakness in Wells Fargo. OK. Now let's move on to the heavyweight—Goldman Sachs. These stocks are currently rising in the premarket. The CEO says that despite the challenging operating environment, clients continue to turn to Goldman Sachs . They said: "Despite the current complexity and uncertainty in the world, we remain well-positioned to continue working for our clients." Wall Street seems to like it.
Yes, that's right. Excellent earnings per share: $171 versus the expected slightly more than $13. This was a 30% increase over the forecast for net profit. Revenue was 18.68 billion versus the 16.71 billion expected. This was almost a 12% overrun. Investment banking fees amounted to 3.22 billion. Equity income grew 72% year-on-year to 6.44 billion.
It was very solid. And remember, CEO Solomon actually stated that fixed income trading revenue might be lower than expected. It decreased by approximately 1% year-on-year . It amounted to 3.47 billion. So it's not as big a miss from the forecast as Wall Street expected. The key takeaway is strong earnings per share and revenue performance, supported by $9.91 billion in combined equity and fixed income revenue.
If you look at the revenue from global banking and markets, it was 13.46 billion. This is higher than the 11.8 billion that was expected. The shares, as I mentioned, were a big positive surprise . And assets under management amounted to $4.1 trillion, which is roughly in line with estimates. Expenses reflect higher transaction-related costs and compensation payments to date .
Consulting revenues exceeded expectations by a couple of hundred million, and investment banking revenues amounted to 3.23 billion. This was higher than the expected 2.86 billion. So, we have an excess of indicators in all areas of Goldman Sachs. It's no wonder we're seeing stocks rise. Remember, this action was in the territory of a "bear " market.
Falling 23% from recent all-time highs over the past few months. So maybe the bar was a little lower for Goldman Sachs, but the results exceeded expectations. . A new problem has emerged in the AI boom . It's not just about how powerful this technology can become . Now investors are asking a much more important question.
Who will pay for all this? CoreWeave , SK Hynix, Micron, Dell and Oracle faced strong selling pressure. CoreWeave alone has reportedly lost about 30% in just six or seven trading days, and the fear is spreading throughout the semiconductor industry. But that's what makes this situation so interesting.
The demand for AI computing still remains high. Companies are still making deals. Data centers continue to be built. So why are investors suddenly rushing to the exits? The answer can come down to one thing: the cost of capital. Let's figure it out. For years, investors have focused on how quickly demand for AI will grow .
Now they are asking whether the companies building this infrastructure will be able to raise enough funds for further growth . Building an AI data center costs billions of dollars. Companies need advanced chips, servers, power, and cooling systems. The huge amount of memory and a significant portion of data costs depend on borrowed funds.
When interest rates rise or lenders become more cautious, these projects become more difficult to finance. Even if demand is high, the business case may become less attractive. This is exactly the fear that has gripped the market now. Take CoreWeave for example. The company continues to sign large contracts, but investors are looking beyond them.
They want to know how the company will finance its future expansion. The contract may guarantee years of profit, but the company still needs money to build the infrastructure to provide these services. If financing becomes too expensive, profits could come under pressure. And if investors believe future borrowing will become more difficult, stocks could fall even as the business continues to attract customers.
And that's why this sale is n't just about demand. It's also a question of whether the money needed to support this demand will remain affordable, and CoreWeave isn't the only company facing this issue. Oracle has also come under pressure due to concerns about its capital needs and debt. This is about $15 billion in debt for CoreWeave and about $167 billion for Oracle.
These numbers help explain why investors pay so much attention to the cost of borrowing. Higher interest rates do not affect every company equally. A company with strong cash flow and a healthy balance sheet can handle them. A company that is forced to borrow heavily to finance its growth has a much greater challenge.
But there is another side to this story. Loans backed by Nvidia chips are reportedly holding up better than some publicly traded debt related to AI infrastructure. This is important as lenders try to understand the value of the equipment behind these loans. If the chips retain their value for years, they could potentially serve as useful collateral.
This would help companies raise money by securing equipment they already own. But there is a nuance. AI chips do not exist in a simple standard market, where each piece of equipment has the same cost. Different companies build and operate their data centers differently. Chips can be used for a variety of contracts, and their value depends on how much useful work they can continue to do.
There is also a risk that new technologies will make older chips less competitive . So the real question is not simply whether a used chip has a certain value. And in how much it will retain its value, how easily it can be returned, and whether creditors are willing to accept it as collateral. If this funding model works, it could support further spending on AI infrastructure.
If not, the entire system may face greater pressure. Now this brings us to a broader discussion. Is the AI boom exhausting demand , or is it facing a funding problem? These are two very different situations. If demand falls, companies that supply chips and servers could face a decline in orders . But if demand remains high and financing becomes more expensive, the pressure could fall more heavily on companies with weak balance sheets and large borrowing needs.
This breakdown is extremely important, as some companies that have been sold off have very different financial situations. CoreWeave and Oracle are in dire need of funding. Nvidia, Broadcom, and Micron have different business models and funding profiles . Their shares may still fall during a broad market sell-off , but that doesn't mean their core business faces the same risks.
Investors need to distinguish between companies that can finance their growth and those that depend on constant access to external capital. And now the market is trying to figure out where that line is.
Comments 0