article

U.S. Stock "Tech Sisters" See Declines

The stock prices of the U.S. "Tech Sisters" reached their peaks within the year between mid-June and mid-July, with Microsoft's share price exceeding $460 per share, Apple's over $230 per share, NVIDIA's over $135 per share, Alphabet's Class A over $190 per share, Meta's over $530 per share, Amazon's over $200 per share, and Tesla's over $260 per share. Last Friday, except for a slight increase in Apple's stock, the share prices of the other six giants all fell.

On Monday of this week, the share prices of the "Sisters" experienced a collective decline. As of the time of writing, the intraday share prices were Microsoft at $394.99, Apple at $203.1, NVIDIA at $95.38, Alphabet's Class A at $157.21, Meta at $455.98, Amazon at $158.8, and Tesla at $196.31.

Behind the sharp decline in many U.S. stocks is the rise in the U.S. unemployment rate in July and a cooling of the job market that exceeded market expectations. Deutsche Bank strategist Jim Reid stated that the market was tense before Friday, and the weak employment data did indeed exacerbate significant fluctuations globally, "From today on, the most important thing is to stay vigilant."

Advertisement

Buffett has been significantly reducing his holdings in Apple. According to the latest financial report, Berkshire Hathaway, under Buffett, reduced its Apple shares by nearly 50% in the second quarter of this year, while Berkshire's cash reserves reached a new high of $276.9 billion. Behind this, Buffett mentioned at the shareholders' meeting in May this year that he expected the company's cash reserves to increase, and said, "We want to spend this money, but if we spend it, we must find the right target. We hope that the risk can be small enough, and the return can also be large enough."

In addition, looking at the operational situation of U.S. tech companies, several tech giants are increasing their investment in AI, and investors have some concerns about when these investments will be converted into revenue and profits.

According to the second quarter financial report of Alphabet, Google's parent company, released in July, the company's capital expenditure on purchasing technology infrastructure reached $13.2 billion in the quarter, with the investment aim of supporting business growth, especially for AI products and services, which is expected to be higher than in 2023. The company is expected to have a capital expenditure of no less than $12 billion each quarter this year.

Meta's latest financial report states that the company expects its AI strategy to require increased investment in infrastructure and personnel. The cash used for purchasing servers, data centers, and network infrastructure and other properties and equipment in the first half of the year reached $14.57 billion, and it is expected that the capital expenditure in 2024 will reach $37 billion to $40 billion, with capital expenditure also expected to increase significantly in 2025.

Microsoft's latest quarterly capital expenditure, including financing leases, reached $19 billion, with cloud and AI-related expenditures accounting for the vast majority of it, with nearly half invested in infrastructure construction, and the company also plans to continue increasing capital expenditure. "When there are signals of demand (growth), we drive most of the capital expenditure based on this. Data centers, land, etc., are long-term assets, and we need to effectively use capital expenditure to build long-term assets," Microsoft's management said in a conference call.

Apple CEO Tim Cook stated during the earnings call that as time goes on, the company has increased its research and development efforts, and Apple has been investing in AI and machine learning. In terms of capital expenditure, it can be expected that investments will increase year by year. Amazon's CFO Brian Olsavsky stated in the call that Amazon's capital expenditure in the first half of the year was $30.5 billion, and the capital investment in the second half of the year will be even higher, mostly used to meet the demand for the growth of Amazon Web Services (AWS) infrastructure.Has AI Investment Already Created a "Bubble"?

A significant influx of capital expenditure in a short period has affected the financial performance of some tech companies. Microsoft executives stated during an earnings call that the gross margin of Microsoft Cloud is around 70%, with a year-on-year decline mainly due to the expansion of AI infrastructure scale. Intel, on the other hand, reported a net loss of $1.6 billion in the second quarter of the fiscal year 2024 and plans to lay off approximately 15,000 employees. Intel's management indicated that the profitability in the second quarter was lower than expected, partly because the company decided to accelerate the production of Ultra AI PC CPUs.

During the earnings call, investors and institutions were particularly concerned about when they would see a return on their capital expenditures. Some U.S. tech giants emphasized the lag in returns and the necessity of increasing investments in the AI sector at present.

"Including land acquisition, data center construction, and financing leases, these projects may take 15 years or even longer to monetize," Microsoft's management pointed out. Amazon CEO Andy Jassy stated that generative AI components are still in their early stages, and although the company has already generated billions of dollars in revenue from the AI sector, it is still early days.

Alphabet executives emphasized that the risk of underinvestment in AI far outweighs the risk of overinvestment. Intel's management expressed that investments made to define and drive the AI PC category will put short-term pressure on profits, but the company believes it is worth it, as current investments will lead to significant growth in the coming years.

The lag in returns also implies uncertainty. Currently, generative AI applications are still in their infancy, and many cloud giants are investing more in infrastructure such as data centers and server purchases for large model training and cloud inference. Reporters have learned from several server industry insiders that this year, the most noticeable growth in server sales is in AI servers for data centers, while the most significant increase in procurement volume is from overseas tech giants.

However, to get consumers to ultimately pay for AI services, it is necessary to make the cost of using AI more affordable and to implement AI applications on commonly used devices such as computers and smartphones.

"The price cannot be too high, large models need to be miniaturized onto devices, and the issue of high energy consumption must be addressed. These will be key to whether AI can enter terminal devices and attract more consumers to use it. Currently, there is mainly an expectation for better development results of AI PCs and AI smartphones in the first half of next year; it is still difficult to say how much of a dividend AI will bring in the second half of this year," an AI industry chain insider told reporters recently. Another AI industry chain insider told reporters that using AI services in the cloud is still too expensive, and in the future, there needs to be a democratization of computing devices similar to home computers to quickly boost demand.

There are reports that hedge fund Elliott Management recently told investors that large tech companies, especially Nvidia, are in a "bubble." The fund questions whether large tech companies will continue to buy Nvidia's GPUs on a large scale, and that AI is overhyped, with many applications not yet ready for AI. Many assumed uses of AI will never be cost-effective, will not play a real role, and will consume too much energy.

Goldman Sachs analyst Jim Covello recently stated that companies spending hundreds of billions of dollars in the AI sector will not trigger the next economic revolution. Most historical technological transformations have replaced expensive solutions with cheaper ones, and replacing low-wage jobs with costly technology (AI) is completely contrary to the technological transformations he has witnessed before. The cost must be significantly reduced to make it affordable for the general public to use AI to perform tasks."If investors start investing in (technology giants) now and expect returns in 10 to 15 years, this is venture capital investment, not public company investment. For public companies, we hope to see returns on our investment in a shorter time frame. Therefore, (the increase in capital expenditure by technology giants for AI) has caused unease, as there are currently no visible applications that can generate revenue to justify the rationality of these investments," said D.A. Davidson analyst Gil Luria recently.

There are still institutions that are optimistic about the prospects of AI. Morgan Stanley recently included NVIDIA in its preferred stock list, with the institution's analysts stating that they are optimistic about NVIDIA's long-term growth potential, and demand for Blackwell chips will be a catalyst for the rise in NVIDIA's stock price.