In this round of sell-off of U.S. technology stocks, Nvidia, as the leader of GPUs, has also pulled back a lot, and this week's stock price is more like a roller coaster, falling 7% on Tuesday, then rising nearly 13% on Wednesday, and falling 6% on Thursday... Shares have fallen more than 19% since their July highs.
Morgan Stanley ignored Nvidia's recent decline in stock price, and released a research report on July 31 to be optimistic about Nvidia, and pointed out that due to Nvidia's current strong sales of chip products, and the prospect of new products in the future, as well as the negative impact of the market's excessive exaggeration of industry competition will fade over time, Nvidia's stock price correction will provide a good entry point. Still strongly bullish on Nvidia's future stock price performance and maintained an "overweight" rating. Nvidia shares jumped nearly 13% at the close of trading on Wednesday on the news.
So why does Da Mo continue to be optimistic about Nvidia?
Because Da Mo believes that the current market concerns about Nvidia are mainly focused on several aspects such as excessive capital expenditure in the AI industry and fierce competition in the industry, and these negative impacts will be weakened over time, after all, Nvidia's GPU product advantages are still there.
The reasons why Da Mo is optimistic about Nvidia:
1. In the future, the demand for NVIDIA GPU products will gradually shift from Hopper to Blackwell, and during the transition period, the sales of H100 and H200 GPU products with Hopper architecture will remain strong, which can support Nvidia's revenue.
It is clear that Hopper GPU products are at the end of their cycle, and with technological advancements and product upgrades, market demand will shift from Hopper-based products to the next generation of Blackwell architecture products.
As demand shifts for products, supply chain constraints change: from a lack of GPU capacity to a lack of supply of silicon materials (i.e., the supply of raw materials can become a new bottleneck).
2. The enthusiasm of large technology companies for new GPU products based on the Blackwell architecture continues to rise, making NVIDIA's future revenue prospects clearer.
NVIDIA's new Blackwell product offers significantly improved inference performance over its predecessor, increasing customer appeal.
While the complexity of the Blackwell and combo racks may pose supply challenges, Nvidia's ramp-up appears to be progressing well.
We expect the first Blackwell shipments to be significant by October, but Hopper-based GPUs will remain a major source of revenue for NVIDIA by early 2025.
Da Mo: The recent decline in Nvidia's stock price may be related to the following risks, but don't worry too much!
Da Mo is very optimistic about Nvidia, and pointed out that the recent decline in Nvidia's stock price may be related to the following risks, but the negative impact of these risks will diminish over time.
The first risk: tech giants significantly increase AI capital spending, but are unable to recoup a proportionate return on investment from AI investment
We've seen AI capex continue to rise everywhere, but Wall Street has recently been unhappy with the tech giants' massive increase in quarterly capex on AI, triggering a recent sell-off in U.S. tech stocks.
But to understand that big tech companies are facing a variety of land and grid power constraints as they ramp up their AI investments, capex is non-linear.
Our surveys generally show that tech companies want to deploy GPUs as soon as possible, and demand for Nvidia's H100 products remains stable even with the upcoming launch of the new product Blackwell, which may have been overly concerned about the need for H100 products by tech companies.
The second risk: the competitive ecosystem
Da Mo pointed out that artificial intelligence is a huge market, and Nvidia's GPU products alone cannot meet the needs of the entire market, so Amazon wants to build its own chips, and Apple is also considering using Google's TPU to complete the key work of training large models, and AMD is also carving up the GPU market, these competitions have put pressure on Nvidia's stock.
Fortunately, we've heard that many tech companies that have invested in custom chips or alternatives are back in the arms of Nvidia, which is still at the center of the AI race, and we don't think that's going to change for a while.
The third risk: supply chain issues
Although Nvidia is currently facing some supply chain concerns, such as HBM3e production capacity and heat dissipation of GB200 rack products, Nvidia is actively addressing the problem, and there are positive signs that Blackwell products will accelerate mass production in the second half of the year.
At the same time, Da Mo pointed out another point to be wary of: some of Nvidia's foundry partners (ODMs) in Taiwan may be too optimistic about the future demand forecast for GB200 products.
Some ODMs predict that the demand for GB200 product racks may be as high as 5~90,000 next year, which may be a bit optimistic.
Based on an average of 54 GPUs integrated into each GB200 rack, this would cost $150 billion, assuming a market demand of 70,000 racks and a price of about $40,000 per rack ($35,000 for GPUs and $5,000 for CPUs and switches).
Therefore, we expect that next year, Nvidia's data center business revenue is unlikely to reach such an optimistic level of $300 billion, but closer to $164 billion. If you exclude the revenue from Mellanox and the gaming business, the revenue from processors is about $145 billion.
During this period, the biggest risk remains the fluctuation of revenue figures. However, in the short term, the data will still support Nvidia's stock price.
Fourth risk: Macroeconomic environment concerns and compressed market valuations
Da Mo pointed out that the economic environment may be the most difficult factor to ignore:
Because high capex for cloud computing and GPUs relies on a strong risk appetite environment, a weakening global economy will certainly affect spending. At that time, investors' valuations of companies may become more stringent, which will affect investors' willingness to invest in risky assets.
Finally, Da Mo pointed out that there are many catalysts that could boost Nvidia's stock price in the future, such as upward revisions to earnings forecasts or earnings expectations, more certain delivery times for Blackwell products, Nvidia's strong response to competition concerns in the market, and the fact that AI is a topic in the semiconductor industry that makes Nvidia's business more visible.
This article is from Wall Street News, welcome to download the APP to see more