Cisco 24 years ago? Bubble theory has resurfaced since Nvidia reached the top, but Huang has set his sights on governments

National Business Daily

2024-06-19 16:17Published on the official account of Sichuan Daily Economic News

Reporter: Cai Ding Editor: Lan Suying

On Tuesday, June 18, Eastern time, Nvidia closed up 3.6%, with a market capitalization of $3.34 trillion, surpassing Microsoft to become the "world's most valuable company" for the first time.

Cisco 24 years ago? Bubble theory has resurfaced since Nvidia reached the top, but Huang has set his sights on governments

Image source: Google Finance

As tech companies like OpenAI, Microsoft, and Alphabet race to snap up GPUs, Nvidia's performance has also skyrocketed. At present, Nvidia's market share in the AI chip market for data centers has reached nearly 80%.

However, being the "world's largest company by market capitalization" doesn't mean Nvidia doesn't have a vision. The rapid growth in the early stage has led more and more people to start worrying about whether this growth can be sustained. If global AI spending shows signs of slowing, optimism in the sector could die with it.

Moreover, on the day that Nvidia's market value peaked, some mainstream financial media in the United States compared it with Cisco, a leading company during the Internet bubble at the beginning of this century, saying that the two have some similarities and warn investors to beware of the bubble behind the boom.

How to find the next growth point is crucial for NVIDIA. The National Business Daily noted that earlier this year, Huang had already targeted the AI "national team" and pushed for the "Sovereign AI" strategy. In this year's quarterly earnings call, Nvidia executives mentioned that the "sovereign AI" business is expected to bring in nearly $10 billion in revenue in 2024.

$3.34 trillion! Nvidia's market capitalization topped the world's No. 1 market

On June 18, Eastern time, Nvidia's market capitalization reached $3.34 trillion, surpassing Microsoft (with a market capitalization of $3.32 trillion) to become the "world's most valuable company" for the first time. This is the first time that a company other than Microsoft and Apple has won the title since Amazon briefly topped the summit in February 2019.

Top 10 companies by market capitalization Image source: companiesmarketcap

Under the wave of AI set off by ChatGPT, NVIDIA's ultra-high computing power GPUs have been in short supply. Nvidia executives recently said that the demand for the latest generation of AI chips, Blackwell, is likely to be just as strong and "continue into next year."

As the demand for chips continues to rise, NVIDIA's performance has also shown explosive growth. According to the data, as of the first quarter of this year, Nvidia's net profit has exceeded analysts' expectations for six consecutive quarters, and its revenue has exceeded expectations for seven consecutive quarters.

Analysts are also looking forward to Nvidia's prospects. Like Nvidia CEO Jensen Huang, they also believe that AI is the foundation of a new industrial revolution, and Nvidia will play a central role in this revolution.

Daniel Ives, managing director and senior equity analyst at Wedbush Securities, said in an email to the Daily Economic News that he believes in the coming year, Nvidia, Apple and Microsoft will be the focus of the $4 trillion market value of the technology industry. With the advancement of the Fourth Industrial Revolution, the demand for high-performance chips from enterprises and consumers is growing, and in essence, Nvidia's GPU chips are the new gold or oil in the technology field.

According to the London Stock Exchange Group, Nvidia has become the largest company on Wall Street by trading volume, with an average daily trading volume of $50 billion, while Apple, Microsoft and Tesla have an average daily trading volume of about $10 billion. Currently, Nvidia accounts for about 16% of all transactions in the S&P 500 index.

Nvidia is the largest company on Wall Street by trading volume Credit: Reuters

The optimism in the secondary market has also allowed Nvidia's valuation to grow significantly in a short period of time. In February this year, Nvidia's market capitalization rose from $1 trillion to $2 trillion in 9 months, and from $2 trillion to $3 trillion, Nvidia only took more than 3 months.

Was it Cisco 24 years ago? Nvidia faces two real-world challenges

However, such a large increase in Nvidia's short period of time has also made more and more market watchers begin to question the rationality and sustainability of its valuation.

On the day Nvidia's market capitalization peaked, the Wall Street Journal published an article saying that Nvidia was leading a technology boom, reminiscent of the Internet boom at the beginning of this century.

Image source: Screenshot of the Wall Street Journal report

The last time a major computing infrastructure provider became the world's most valuable company was in March 2000. At the height of the Internet boom, network equipment company Cisco also surpassed Microsoft to reach the top.

According to the report, John Chambers, who served as the CEO of Cisco during the Internet boom, believes that Nvidia and Cisco at that time do have some similarities, but the development of AI is different from the revolution led by the Internet and cloud computing. "In terms of the size of the market opportunity, the impact is equivalent to the sum of the Internet and cloud computing. The speed of change is different, the market size is different, and the stage at which the company reaches its maximum market capitalization is also different. Chambers said.

In fact, this is not the first time that the mainstream media has compared Nvidia to Cisco during the dot-com bubble at the beginning of this century. In February, the Financial Times drew an analogy between Nvidia and Cisco, warning investors and institutions to remember history.

At that time, hundreds of millions of investors around the world expected the Internet to change the world, and the demand for network hardware such as servers and related routers also increased dramatically, and telecom companies and hardware suppliers became the biggest winners in the capital market, and the stock prices of telecom equipment stocks such as Cisco peaked before the dot-com bubble burst in 2000. Among them, Cisco's stock price has skyrocketed more than 30 times in just a few years.

However, the telecom industry crashed earlier than expected due to the Internet speculation boom – it took the entire industry just four years to go from boom to bust and much faster than the Internet could change people's lives. By 2002, two years after the dot-com bubble burst, excess supply had bankrupted more than 20 telecommunications groups, as well as a sharp sell-off in the shares of other companies that had not.

For NVIDIA, the real challenge is that if the technology manufacturers cannot achieve tangible performance growth through large models in the next 1~2 years, then NVIDIA's performance growth will also lack a solid foundation.

The Wall Street Journal quoted Sequoia Capital as saying in March that Nvidia's chips have received about $50 billion in investment since the release of ChatGPT sparked an AI boom, but the sales of generative AI startups are only $3 billion.

According to the Financial Times, another big challenge for Nvidia is Moore's Law, which means that high-end computing chips will soon be commercialized.

For example, the old 40-nanometer chips used in household appliances are no longer in short supply, but when they were first launched in 2008, they were scarce cutting-edge products. With the depreciation of capital equipment, the price of such chips has fallen sharply in a short period of time.

The speed of chip development and manufacturing is accelerating every year, and the efficiency of software is improving every year. It took only two years for chips to be upgraded from the 7nm process to the 5nm process used by NVIDIA's latest chips. This rapid technological advancement means that companies in the industry are likely to spend much less on chips in the future than they expect today.

"Sovereign AI" as a new growth driver? Revenue is expected to approach $10 billion this year

So how should NVIDIA maintain high growth, and where is the next growth point?

At present, in addition to GPU sales, Nvidia is vigorously promoting the "sovereign AI" strategy. At the 2024 World Government Summit in February this year, Nvidia CEO Jensen Huang bluntly said that each country should build its own "sovereign AI".

Image source: Nvidia's official website

The reporter of "Daily Economic News" noted that this concept is not the first of its kind - in 2020, the European Parliament passed a proposal on how to better regulate AI, emphasizing the ethical principles and legal obligations that should be followed when developing, deploying and using AI. In addition, the EU is also concerned about data sovereignty issues, and in 2018 implemented the EU General Data Protection Regulation, which regulates all businesses operating in the EU. This is roughly the prototype of the concept of "sovereign AI".

On the official website, Nvidia has a clear explanation of "sovereign AI": "sovereign AI" covers both physical and data infrastructure. The latter includes autonomous foundational models, such as large language models, developed by local teams and trained on local datasets to promote inclusivity in specific dialects, cultures, and practices.

Countries in Asia, the Middle East, Europe and the Americas are now investing billions of dollars in new AI computing facilities in their countries, opening up channels for Nvidia and other technology companies to boost sales growth. Governments are increasing their budgets and providing other incentives to encourage their companies and multinationals to build new data centers and retrofit old ones with specialized computer chips, mostly from Nvidia.

A press release released on Nvidia's official website also shows that Nvidia's "sovereign AI" customers include governments in Singapore, Japan, France, Italy, India and other countries and regions.

According to media reports, Singapore's National Supercomputing Centre is being upgraded with Nvidia's latest AI chips, while state-owned Singtel is working with Nvidia to push to expand its data center footprint in Southeast Asia. Canada also pledged $1.5 billion last month as part of its sovereign computing strategy to support its start-ups and researchers. After Huang's visit to Japan, Japan said it would spend about $740 million this year to develop its AI computing capabilities in the country.

Nvidia executives disclosed in the first quarter 2024 earnings call that the "sovereign AI" business is expected to bring in close to $10 billion in revenue in 2024, although the business has not yet generated revenue in 2023.

View original image 43K