NVIDIA's financial report is not important? Morgan Stanley: What we should really pay attention to is whether Black Myth: Wukong will be delayed
Morgan Stanley believes that what truly affects NVIDIA's stock price is not this week's financial report, but whether Blackwell will be delayed. After investigating and researching the chip industry chain, Morgan Stanley optimistically stated that the current situation in the artificial intelligence chip market is very good
NVIDIA's Q2 quarterly financial report is about to be announced. The financial performance of this chip giant will not only directly affect the performance of the technology industry, but may also stir the pulse of the entire financial market.
In its report on the 25th, Morgan Stanley expressed optimism about NVIDIA's performance, stating, "We expect this to be a strong quarter, likely to exceed higher expectations." The market's general expectation is that NVIDIA's Q2 revenue will reach $28.68 billion.
Morgan Stanley pointed out that if NVIDIA's revenue exceeds expectations, the prices of AI-related stocks may have an upside potential of 3-15%. Conversely, if it falls short of expectations, the entire AI stock group may have a downside potential of 5-10%, and the priority of stock selection may reverse.
Morgan Stanley also stated that what truly affects the stock is not this week's financial report, but whether Blackwell will be delayed. After investigating and researching the chip industry chain, Morgan Stanley optimistically stated that the current situation of the artificial intelligence chip market is very good.
What really matters is whether Blackwell will be delayed
Currently, the market generally expects NVIDIA's Q2 revenue to double year-on-year, reaching $28.68 billion. Morgan Stanley believes that whether the revenue for the third quarter meets the market's expectation of $31.5 billion, or some analysts' expectations of over $33 billion, may not have a significant impact on the stock.
What truly affects the stock is whether the company can alleviate investors' concerns about the potential delay of Blackwell due to chip design adjustments.
Regarding the delay concerns, Morgan Stanley stated that the possibility of such a delay cannot be ruled out. However, through investigations of the supply chain and observations of strong procurement demand and increased capital expenditures by cloud computing companies, Morgan Stanley concluded: the current situation of the artificial intelligence chip market is very good.
Morgan Stanley predicts that the delay will not affect the full-year prospects of upstream companies (such as foundries, CoWoS, and HBM).
NVIDIA remains the most sought-after silicon chip supplier in the market, with major data center companies increasing capital expenditures to cope with price increases.
The money NVIDIA makes is the money spent by mega-enterprises, and all these capital expenditures flow directly to NVIDIA's supply chain, which is also where the revenue is concentrated.
Morgan Stanley pointed out that despite NVIDIA's rapid revenue growth, investors are still concerned about the sustainability of this growth and return on investment (ROI). However, thanks to the continued high capital expenditures of large U.S. data center companies, NVIDIA's supply chain revenue remains concentrated.
Morgan Stanley: NVIDIA will exceed expectations with 98% growth, not recommended to give up AI-related stocks
Morgan Stanley suggests that even if the third-quarter financial report is disappointing, "do not give up on AI-related stocks." The market generally expects NVIDIA's revenue to grow by 98% this year, and Morgan Stanley believes that this quarter may once again exceed expectations and raise guidance.
Currently, investors are paying a forward 12-month price-to-earnings ratio of 38 times for NVIDIA, slightly higher or in line with market expectations for its long-term 10-year average
For other AI-related companies, the average forward P/E ratio is 30 times. Therefore, it is difficult to say that these AI stocks have entered the overvalued territory.
In other words, although NVIDIA's high valuation and earnings expectations are concerning, there is currently no clear sign that AI-related stocks are overvalued. The key is to combine earnings with valuation, and for most companies in the AI supply chain, current earnings can still support these valuation multiples.
Morgan Stanley believes that although the valuation of AI stocks has been readjusted, it is far from reaching bubble levels, and some pullbacks in stock prices during the price discovery phase are normal.
The market consensus expected returns of AI stocks are still rising, demonstrating a stronger resilience in the U.S. economy. Even in the event of an economic slowdown, companies still need to invest in AI technology, otherwise they will worry about falling behind or even being eliminated.
Data from the supply chain supports their view that strong demand will continue into next year. Specifically:
Forecasts for the GB200 rack system remain unchanged, but large-scale production may start in the later part of 2025.
TSMC has no production issues, with third-quarter revenue guidance at $22.4-23.2 billion (a 9.5% increase quarter-on-quarter) and full-year growth guidance slightly above 20% year-on-year.
Hyper-scale data center operators are more active in AI server procurement, but Microsoft has reduced its AI allocation in capital expenditures. This is consistent with Aspeed's observation, as Aspeed noted that some U.S. cloud service providers purchased more baseboard management controllers for traditional servers in the first half of 2024.
Meanwhile, Chinese cloud service providers are showing positive growth in capital expenditures.
Tencent's operating capital expenditures in the second quarter of 2024 were 7.2 billion RMB, a 9% increase quarter-on-quarter and a 140% increase year-on-year, mainly driven by investments in GPU and CPU servers.
Alibaba's capital expenditures in the second quarter of 2024 also reached 12 billion RMB, a 99% increase year-on-year and a 9% increase quarter-on-quarter, mainly due to very strong demand and backlog orders