저자: JeremySiegel · 외국인
요약 보기
시겔은 임금 상승률(3.2%)이 CPI 예상치(3.4%)보다 낮아 노동자 구매력이 악화되고 있다고 지적하며, 생산성 둔화를 문제로 봤다. AI는 GPT 출시(2022.11) 이후 연평균 생산성 2.5%로 15년 평균보다 낫다고 평가했다. 엔비디아 실적은 EPS 2.22달러(예상 2.09), 매출 962억달러(예상 923억)로 양호했고 다음 분기 가이던스도 상회했으나 주가는 약 2.8% 하락했다. 다만 AI 투자가 부채로 조달되는 순환 금융 구조에 대한 비판도 언급했다.
다듬은 글
Good morning. Good morning, Melissa. Is there anything about the structure of this market rally that worries you? Well, for example, the fact that it is so dependent on AI, or the fact that AI spending is now largely financed by debt—something like that? Well, let me tell you what I'm looking at. You know, the most important variable for the economy is actually not inflation or employment, but wages adjusted for inflation .
After all, this is real purchasing power. And, honestly, the most disappointing data I saw on Friday was the wage growth figures. And they grew by only 3.2%. Now on Wednesday we will get data on the Consumer Price Index (CPI). It is expected to grow by 3.4%. So, what does this mean? Workers are losing the fight against inflation.
What is the main source of workers winning in the fight against inflation? This is productivity growth. And, frankly, the productivity growth over the last three quarters has been very disappointing. We need to speed up this process. There are some positive signals for this quarter, but the figures for the last three quarters have been below the 15-year average.
And this is one of the reasons why , in my opinion, many people are unhappy with the state of the economy: they are not ahead of inflation. So we need to see how AI really works to increase the real wages of ordinary people. I'm very interested in my good friend Kevin's opinion. I heard you're going to be interviewing him in a few minutes.
I would love to hear his perspective, but I think that's what we need to see in this economy to turn things around for the better and restore people's confidence. Do you think we are now in a transition phase where we are spending on AI, AI is causing inflation in the economy, and workers are not yet benefiting from it ?
That is, we are in a period of transition. Looking at it more broadly, GPT was introduced in November 2022. The average productivity growth since then—and that's almost four years— has actually been 2.5%, and that's not bad, it's better than the last 15 years. It's not as good as the "golden years" of the 1950s and 60s, when we had over 3%, but it's better, although we've slowed down a bit in the last three quarters.
Maybe it's about oil or other factors, but we need to get back on track. Now for something positive... Let's take a look at Nvidia. The Super Bowl of financial reports is upon us. We see EPS at $2.22. This is better than the expected $2.09. Total revenue amounted to 96.2 billion. This is slightly better than the expected 92.3 billion.
The most important revenue from data centers is 89.0 billion, compared to the expected 85.86. Nvidia's gross margin for the quarter was 75%. They also provide some predictions. For the third quarter, the current next quarter, they are forecasting revenue of 108 billion plus or minus 2%. This forecast is higher than the consensus analysis.
They also report that the Blackwell architecture is now in full production. And we see Jensen Huang saying that AI has reached its tipping point. He also used similar expressions. They are also declaring their next quarterly cash dividend of 25 cents per share on October 1st. And for this quarter, they are forecasting revenue of 105.84 to 110.16 billion.
The market expected 105.15. So, it's above that range. They have excess indicators. They are having better sales than expected. They have better earnings per share (EPS). But for some reason, Alex, I knew I would look at the screen and see red, and it was red. The stock fell about 2.8%, but that's not as big a move as was expected.
Obviously, many people are waiting to hear what Jensen Huang himself has to say . But let's look at a few more details about income distribution. So let's quickly go through these numbers before we start the discussion. Revenue from computing and networking was 88.3 billion. This is about 4 billion better than expected.
Revenue from AI clouds, industrial and enterprise —this is the new segment for hyperscalers— was $40.31 billion. Oh , we're getting two numbers here. Revenue from hyperscalers is 48.71 billion. That's also about 4 billion better than expected. Revenue from edge computing was 7.2 billion. That's better than 6.61 billion.
They expect gross margin in the current quarter to be slightly lower than this one: between 73.5% and 74.5%. We had 75% this quarter. Operating expenses for this quarter are expected to be around 9 billion. Expectations were slightly lower than this number. So, they say their liabilities have gone from 119 billion last quarter to 279 billion this quarter.
So, Alex, they showed impressive results, but the question is, were they high enough? And now it seems that the sentiment is leaning towards no. It's very difficult to sit here and try to criticize fundamentals. Numbers are numbers, with Nvidia perhaps just unleashing the next phase of the AI boom, and it has less to do with GPUs than most people realize.
The biggest shift happening now is that AI infrastructure is becoming a holistic ecosystem. Nvidia doesn't just sell chips anymore . It provides the computing platform, networks, software and, increasingly, the financial support needed to build these vast AI systems. And that means the next wave of AI spending is getting much bigger.
Hyperscalers continue to expand data centers, and AI labs require more computing power. And companies are starting to move from AI experiments to real products. But beneath all this capital lies a bottleneck . This is capital. Building an AI data center can require billions of dollars before it starts generating significant revenue.
That's why Nvidia's willingness to invest in AI companies and support the funding structure has attracted so much attention. Critics see this as circular financing . The "bulls" argument is different. If Nvidia helps an AI company get the infrastructure it needs, that company will eventually make enough profit to pay for the computation.
And Nvidia isn't just creating demand. It accelerates the development of the ecosystem around its own platform. And there is one more thing that investors should not ignore. This is memory. AI is creating a huge demand for high-bandwidth memory . DRAM and data storage. Micron's CEO even claimed that memory has evolved from a basic component to a strategic AI infrastructure.
This created a fascinating supply problem. You can build more data centers, you can order more GPUs, but if critical memory components are missing, the entire system can hit a wall . And Nvidia is still at the heart of this process. Now the company has to prove another important variable. This is its next- generation platform.
If all goes smoothly, Nvidia won't just defend its current leadership in AI. It will seek to extend this advantage to the market is watching for more than just Nvidia's quarterly report. Investors are watching the whole mechanism: demand for AI, data center construction, memory supply, capital availability, customer spending, and Nvidia's ability to turn it all into profit.
The real question is not whether spending on AI continues. This is obvious. The question is how long this infrastructure cycle can continue before supply, financing, competition, or customer profitability become a deterrent. And this is the part of Nvidia's story that can determine the future direction of the entire artificial intelligence sector.
Comments 0