laitimes

Artificial intelligence engine "goes out"

author:21st Century Business Herald
Artificial intelligence engine "goes out"
Artificial intelligence engine "goes out"

Author丨Wu Bin

Editor丨Li Yingliang

Source丨AI

After a round of "anti-sky" rally, the artificial intelligence boom and the U.S. stock rally are showing signs of weakness, and the engine of yesteryear is no longer standing.

On April 9, Nvidia fell about 5% intraday and finally closed down more than 2% to $853.54, and its stock price has fallen more than $100 from its all-time high. Since the AI boom peaked at the end of February, AI concept stocks in the U.S. stock market have almost stood still in the past six weeks. Entering April, the Nasdaq also showed signs of "not moving", and has fallen overall so far this month.

While the performance of technology stocks is gradually returning to normal, cyclicals are moving higher on the back of economic data, and the balance of the rally in US stocks has become relatively balanced. Zhao Wei, chief economist of Guojin Securities, told the 21st Century Business Herald reporter that after 2021, the global manufacturing industry has experienced a two-year downward trend, and since the second half of 2023, the global manufacturing industry has bottomed out and rebounded, and the repair curve has gradually become clear, and the global manufacturing PMI index has rebounded from 48.6 in July 2023 to 50.6 in March 2024, and has stood on the boom and bust line for nearly three consecutive months.

By all indications, the market breadth indicator is improving significantly. Ryan Detrick, chief market strategist at Carson Group, said the key to the bull market was sector rotation, and it has now seen some money rotate away from tech stocks and into some cyclical stocks. The number of S&P 500 constituents, which recently reached a new 52-week high, reached 118, the highest in three years, indicating an improvement in market breadth.

The "blue ocean" gradually becomes the "red ocean"

As the "most important stock on the planet", Nvidia's hegemony is not without risks, and the "blue ocean" that Nvidia traveled earlier seems to have the potential to become a "red ocean".

On April 9, Intel released its latest artificial intelligence chip, Gaudi 3, which is expected to be available to customers on a large scale in the third quarter, and companies including Dell, HP and Supermicro will use the Gaudi 3 chip. Intel claims that the new Gaudi 3 chip will have an average of 50% better inference power, an average 40% better energy efficiency than the Nvidia H100 chip, and run AI models 1.5 times faster than the H100, and that the Gaudi 3 will be roughly on par with Nvidia's latest H200 and even better in some areas.

Das Kamhout, vice president and senior principal engineer of Intel's Cloud and Enterprise Solutions Group, expects the Gaudi 3 to compete with NVIDIA's latest chip, "which we believe is a powerful product based on its competitive price, unique open integrated networking, and standard Ethernet use." Intel CEO Pat Gelsinger also admitted that the earlier version of Gaudi failed to achieve the market share growth that Intel had been hoping for, but he expected the new version to have a greater impact.

Nvidia is facing increasing competition in the field of artificial intelligence. Brian Colello, equity strategist at Morningstar, said no other company has as much dominance in the market as Nvidia, but other tech giants are certainly "fighting their way". In addition, Nvidia could also be hit if the economy slows down, and it could spell trouble for Nvidia if customers like Microsoft scale back their GPU spending, and investors already have high expectations for Nvidia.

There have been some bearish voices in the market at the moment. DA Davidson analyst Gil Luria is skeptical of Nvidia's staggering valuation, believing that Nvidia's huge gains may soon be over, with Nvidia shares expected to fall by up to 20% by the end of the year. Luria has a "Hold" rating and a price target of $620. "Big Nvidia customers like Microsoft and Amazon are stockpiling GPUs, but as they get to the required amounts, tech giants may not continue to buy in large quantities for years to come. ”

Can Nvidia hold on to the "Iron Throne"?

After a year of more than 200% revenue growth, Nvidia's data center business now accounts for more than three-quarters of the company's overall revenue. While Nvidia can't maintain this level of growth forever, Wall Street generally expects revenue to remain strong in the coming years.

Although Nvidia's stock price has recently lost momentum and has fallen more than 12% from its all-time high, many institutions believe that this is a normal correction.

For example, KeyBanc maintained a Buy rating on Nvidia while raising its price target to $1,200 from $1,100, barring anything that could undermine the overall recovery of the semiconductor industry, Nvidia's share price should remain upside and likely to reach its price target in the next year or so.

Zhang Zhuran, chief researcher of Shanghai Zezhaojia Information Technology Co., Ltd., told the 21st Century Business Herald reporter that after the previous surge, it is very normal for Nvidia's stock price to pull back moderately, and the previous investment channels were too crowded, which does not mean that there will be big problems in the artificial intelligence industry. Even after the pullback, Nvidia has risen by 72.36% this year.

In fact, shaking Nvidia is no easy task, and the hegemony is still quite solid. Nvidia currently occupies 80% of the AI chip market and is the absolute leader, and in the past year, Nvidia's GPU has been the preferred AI high-end chip.

Tejas Dessai, a research analyst at Global X ETFs, told the 21st Century Business Herald reporter that we are witnessing an era of change. In the coming years, data centers will replace trillions of dollars worth of chips to match generative AI developments, providing Nvidia with a large addressable market. Nvidia occupies a leading position in the field of high-end artificial intelligence chips and is expected to still occupy the vast majority of the market by 2030.

In addition, NVIDIA's wide range of products greatly increases profitability. In addition to the hardware, Nvidia's CUDA platform has helped cement its position, helping Nvidia defend its market share while building stickiness and loyalty, Dessai said.

In the future, the demand for artificial intelligence chips will also support NVIDIA's performance. According to Dessai's analysis, hyperscale cloud companies, countries and large enterprises clearly have enough demand to buy more chips, and chip demand will remain high for at least three to four quarters in the next three to four quarters before growth shows signs of slowing. In addition, the possibilities of AI chips extend far beyond the data center, and chips will eventually find their way into smartphones, laptops and devices, IoT systems, robotic devices, medical devices, automobiles, and more, leaving room for smaller companies to capture and serve the market.

Looking for "the next batch of Nvidia"

Although Nvidia remains the undisputed hegemon for the foreseeable future, the previous staggering gains mean that the exponential surge in the future may not be repeated. As the impetus for the AI engine wanes, the market is also divided on the future of U.S. stocks.

Wells Fargo believes that the current bull market in U.S. stocks is driven by the long-term growth of artificial intelligence and the concentration of constituent stocks, which has led investors to prioritize growth and discount metrics over traditional valuation metrics. Driven by long-term optimism, investors are increasingly willing to accept higher valuation thresholds and extend investment horizons, and equities do have some upside at current levels.

As a representative of the pessimists, Albert Edwards, global strategist at Société Générale, warned that the US stock market is reflecting a historic bubble, and the current market is feverishly pursuing artificial intelligence, which is very similar to the technology bubble more than two decades ago, and the stock market may see a sharp correction.

Some believe that AI will drive a surge in corporate profits and underpin current high valuations. But Edwards said optimism and corporate earnings expectations have begun to cool, with the percentage of analysts raising earnings estimates falling below 50% of S&P 500 companies.

After the surge in artificial intelligence leaders, the market is desperate to find the "next batch of Nvidias". Anton Levy, chairman of General Atlantic's global technology division, called the current AI investment boom a "big infrastructure" that he believes will lead to greater investment opportunities at the "application layer". "The market size of the application layer opportunity is much larger than that of infrastructure and represents the greatest value creation opportunity for growth investors. The same will be true of the upcoming AI innovation cycle, which will bring unprecedented equity growth investment opportunities over the next decade. ”

Wall Street banks are making moves, with Goldman Sachs Asset Management tapping stocks of AI-powered supply chain component makers, such as cooling systems and power supply stocks, JPMorgan Asset Management favoring traditional electronics makers that are transforming into AI leaders, and Morgan Stanley investment managers betting that AI is reshaping business models in non-tech sectors.

Bank of America analyst Vivek Arya said that while AI deals have benefited chipmakers such as Nvidia and Broadcom, there are also some "second-tier winners" that should rise further, and the rising wave could create a lucrative niche for the next tier of suppliers, with the market on the horizon for a second wave of AI boom.

Looking ahead, Dessai told reporters that the consumption of AI infrastructure bodes well for cloud infrastructure companies as well as AI software manufacturers, and as GPU computing power continues to increase, enterprises are expected to accelerate the adoption of AI software and applications in the future.

SFC

Editor: Liu Xueying, Intern: Huang Lihong

Nvidia enters the game! Humanoid robots have become a "new battlefield" for technology giants

"AI Iron Man" Jensen Huang has evolved Nvidia

OpenAI, "Blowing Up the Field" Again

Artificial intelligence engine "goes out"

Read on