news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

After the market value of AI chip giant Nvidia (NASDAQ: NVDA) plummeted by US$193 billion (approximately RMB 1,396.278 billion) on July 30, Nvidia's U.S. stock rebounded on the 31st, closing up 12.81%. Its market value increased by US$329 billion (approximately RMB 2,378.933 billion) overnight, setting a record high.

This kind of "roller coaster" stock price is even more difficult to predict than Bitcoin. In the early hours of Beijing time this morning, Nvidia's stock price rose much higher than Bitcoin, and the reason behind this was the accumulation of multiple news.

On the news front, on August 1, it was reported that the Biden administration plans to announce a new rule next month that will expand the United States' power to block certain foreign countries from exporting semiconductor manufacturing equipment to Chinese chipmakers. But unnamed sources pointed out that goods from allies such as Japan, the Netherlands and South Korea that export key chip manufacturing equipment will be excluded, limiting the impact of the rule. And Nvidia is expected to sell new AI chips to the Chinese mainland market.

Not only that, according to Bloomberg Billionaires Data, Nvidia CEO Huang Renxun's personal wealth increased by nearly $12 billion on July 31, breaking his own record of $9.6 billion on February 22, thanks to his sale of 240,000 Nvidia shares worth a total of $27 million. In the past month, Huang Renxun has made several large-scale stock sales in July, with a total amount of more than $250 million.

This means that fluctuations of $100 billion have become the norm for Nvidia, the largest company in terms of market value.As of July 30, Nvidia's stock price had plummeted 23% compared to its stock price high a month ago, and its market value had evaporated by nearly $750 billion.

So, why has Nvidia entered into a "roller coaster" situation? Is this AI chip giant about to reach its darkest hour?

Why Nvidia fell sharply: AI companies' performance was poor and competitors' revenue increased significantly

Nearly two years after ChatGPT became popular around the world, Wall Street is worried about the future commercial prospects of AI.

On July 30, the US AI giant Microsoft released its latest financial report. In the second quarter of 2024, the company's total revenue increased by 15% year-on-year to US$64.7 billion; net profit was US$22 billion, a year-on-year increase of 10%.

While Microsoft’s overall sales and profit grew more than expected in its most recent quarter, revenue from its Azure cloud business, which is at the heart of the company’s AI strategy, grew 29%. That was down from 31% growth in the previous quarter and below analysts’ expectations for 30%, according to FactSet.

Influenced by this news, Microsoft's stock price fell 4% after the market closed on the 30th, indicating that Wall Street is disappointed with the core cloud business of Microsoft's AI strategy.

Brad Reback, an analyst at Stifel Financial, said there is a segment of the investment community that is paying close attention to small changes in Azure's business.

In a conference call with analysts, Microsoft Chief Financial Officer Amy Hood said Azure's growth was at the low end of the company's forecast range due to weak demand for non-AI services in some European markets and limitations in AI-related hardware.

"We're constrained in our AI capabilities, and because of that, we ... contracted with third parties to help us," Hood said, describing partnerships Microsoft has forged with other AI vendors. "You do see us investing heavily in building out so that we can get back to a more balanced position."

Not only were the performance of downstream AI application companies weak, but Nvidia's old rival AMD also "assisted" the decline.

On July 30, AMD (NASDAQ: AMD) announced that its second quarter revenue in 2024 reached $5.8 billion, with a gross profit margin of 49%, operating income of $269 million, net income of $265 million, and diluted earnings per share of $0.16. Among them, the data center division's revenue hit a quarterly record high of $2.8 billion, a year-on-year increase of 115% and a month-on-month increase of 21%, mainly due to the substantial increase in AMD Instinct GPU shipments and the strong sales and shipment growth of the fourth-generation AMD EPYC CPU.

“We delivered strong revenue and operating income growth in the second quarter, driven by record revenue in our Data Center business,” said Lisa Su, AMD Chairman and CEO. “Our AI business continues to accelerate, and AMD is well positioned for strong revenue growth in the second half of the year, driven by demand for our Instinct, EPYC and Ryzen processors. The rapid development of generative AI is driving demand for more computing power in all markets, creating significant growth opportunities for us as we integrate leading AI solutions across all our business lines.”

“AMD executed very well in the second quarter, exceeding the midpoint of our guidance, driven by strong growth in both our Data Center and Client business units,” said Jean Hu, AMD executive vice president, chief financial officer and treasurer. “In addition, we expanded gross margins and delivered solid revenue growth while increasing our strategic investments in AI to lay the foundation for future growth.”

A series of factors including poor performance of AI companies, huge increase in competitors' revenue, and even Huang Renxun's frequent selling of Nvidia stock have directly led to Wall Street's concerns about Nvidia itself, with the scale of selling far exceeding the scale of buying.

“The biggest concern is, where is the ROI for all the AI ​​infrastructure spending? The money invested is pretty staggering. Maybe you’ll get your money back in a few years,” said Alec Young, chief investment strategist at Mapsignals. “But I think investors realize that returns take time to materialize and that hyperscale revenues are hurt in the short term by investing so much money.”

Earlier, former US President Trump warned that chip manufacturing is concentrated in Asia, which poses a national security risk to the United States.

This warning directly led to a sharp drop in Nvidia and other semiconductor stocks.

Forbes believes that Nvidia's difficulties on Wall Street this month were not caused by a single event, but the accumulation of several smaller trends. As investors and fund managers cash in profits and adjust their portfolios, it is normal for stocks to fall after a period of continuous rise-Nvidia's stock price is still up 110% so far this year and has risen 610% by the end of 2022, which is indeed a jaw-dropping return for a company of this size. If more people sell than buy, the stock price will of course fall. Nvidia has also been hit by the plunge in semiconductor stocks. As investors prepare for the first interest rate cut since 2020, Nvidia is perhaps the most obvious victim of the market's shift from large technology companies to lagging small businesses.

It is worth mentioning that the share prices of Apple, Microsoft and Google's parent company Alphabet, three other companies with a market value of more than $2 trillion, have all fallen by more than 7% from their all-time highs set earlier this month.

In addition, The Information reported on the evening of July 24 that OpenAI may lose $5 billion this year, and if subsequent financing does not go smoothly, it may face cash shortage in 12 months. Among them, OpenAI recruited 1,500 employees (including on-the-job and outsourced) and invested $1.5 billion, while renting Microsoft servers was as high as $4 billion, while the revenue was only $3.5 billion.

Subsequently, on July 30 alone, several US technology stocks including Nvidia plummeted. Overnight, Nvidia's market value of 1.4 trillion US dollars was wiped out. A new round of AI "bubble theory" has been recognized by more and more investors in the market, causing Nvidia and the entire US stock market to fall into a "correction moment."

Why Nvidia surged: Fed cuts interest rates, tech stocks return

The decline took only one trading day. On July 31, Nvidia's U.S. stock price surged 12%.

Nvidia’s shares rose 150% in the first six months of the year, according to data compiled by Bloomberg. But in July, Nvidia’s stock price plunged, accounting for four of the eight biggest declines in market value.

The volatility comes as investors grapple with a fierce rotation out of high-flying technology stocks and into lagging companies that will benefit from Federal Reserve rate cuts, as well as Wall Street's heavy pressure on technology stocks.

At AMD's earnings conference on Wednesday, AMD CEO Lisa Su revealed that the company's AI chip sales were "higher than expected" and that revenue from the MI300 series chips exceeded $1 billion this quarter.

Based on this performance, Su Zifeng also raised the company's AI chip revenue forecast for this year from the previous $4 billion to $4.5 billion, accounting for about 15% of the overall sales. Since the MI300X has been in short supply since its launch, she expects the chip supply shortage to continue until 2025.

Guoyuan International believes that AMD, as an important potential rival of Nvidia, is working hard to make breakthroughs in AI software in addition to competing with Nvidia in hardware. At present, the demand for AI computing power continues to increase, and Nvidia's supply is obviously difficult to meet all market demand. The company actively deploys the software sector, accumulates technical experience through acquisitions of industry-leading companies to improve shortcomings; at the same time, it keeps a close eye on advanced technologies in the industry. Overall, although it is difficult to compete directly with Nvidia in the short term, the company will continue to benefit from the AI ​​wave.

In terms of performance outlook, AMD expects revenue of approximately $6.7 billion in the third quarter of 2024, with a fluctuation of $300 million. In addition, Microsoft's commitment to invest billions of dollars in its AI infrastructure and Federal Reserve Chairman Jerome Powell's statement on Wednesday that interest rates may be cut in September have driven volatility in the entire market.

After many optimistic predictions, on July 31, Wall Street's concerns about the "bubble" of AI eased, prompting investors to flock back into the technology field and revitalizing the market's enthusiasm for the AI ​​chip industry and Nvidia.

"Microsoft's announcement of increased capital spending, driven primarily by cloud computing and artificial intelligence-related demand, strengthens Nvidia's near-term sales outlook," said Kunjan Sobhani, an analyst at Bloomberg Intelligence.

Matt Maley, chief market strategist at Miller Tabak + Co., said Nvidia's volatility shows how confused investors are right now. They are worried that the huge capital investment in AI will not produce the kind of investment returns that people have been hoping for over the past year. He believes that investors are "worried that the capital investment story may fade." "However, they also know that Nvidia is still performing very well, and every time the stock price rebounds after a big drop, FOMO still plays a role."

Earlier,Huang Renxun said at the SIEPR Economic Summit at Stanford Business School that Nvidia's GPUs are so good that even if competitors such as AMD's AI chips are free, they are not cheap enough to beat Nvidia. Huang Renxun emphasized that in terms of the total cost of ownership (TCO) of AI data centers, the pricing of Nvidia's GPUs is not important, and our TCO is much better than that of our competitors.

In fact, NVIDIA currently has a trump card to defeat other competitors: CUDA.

Most AI companies currently choose Nvidia's CUDA software platform to build their tools. Once these tools are built, they are unlikely to be migrated to run on other companies' chips in the short term. Nvidia has attracted many AI companies through its huge CUDA software platform, established a solid ecosystem and partnerships across dozens of vertical industries, and these companies have difficulty migrating to other platforms due to their huge investments, thus consolidating Nvidia's market position.

Huang Renxun said: "We compete more fiercely than any company on the planet, and in some cases, even Nvidia's customers are competitors. In addition, Nvidia is actively helping customers who are designing alternative AI processors and even revealed to them the upcoming Nvidia chip roadmap."

He admitted that Nvidia operates in a "completely open way" while working with almost everyone else in the industry. People who buy and sell chips consider the price of chips, while people who operate data centers consider operating costs. Of course, companies will be very clear about the total cost of ownership, which basically means Nvidia's advantages, such as deployment time, performance, utilization, and flexibility. Even if a competitor's chip is free, it is not cheap enough.

Huang Renxun emphasized that Nvidia's goal is to "maintain leadership with unparalleled TCO."

In Huang Renxun's view, NVIDIA's core values ​​include two elements. The first is "risk tolerance" and "ability to learn from failure." The second is "intellectual honesty," which means being able to point out the mistakes made by a company or an individual, learn from them, and make quick adjustments. NVIDIA's corporate character is that if you have a good idea and no one has done it before, then give it a try. If you fail, learn and adjust from it. Every failure is a little progress.

"All great ideas start from here." Under the leadership of Huang Renxun, it remains to be seen how NVIDIA will develop in the next AI era.

According to the plan, NVIDIA will announce its second-quarter financial results for fiscal year 2025 on August 28.

(This article was first published on Titanium Media App, author: Lin Zhijia, editor: Hu Runfeng)