NVIDIA is just one step away from "becoming a legend"

Author | Chai Xuchen

Editor | Zhou Zhiyu

Every wave of technological innovation can create a myth of wealth, and the "Leather Jacket Master" Huang Renxun has also caught this wave of prosperity.

On June 5th, Eastern Time, NVIDIA's stock price soared, surpassing a market value of over $3 trillion. That day, it also surpassed Apple to become the world's second highest market value company.

The next day, NVIDIA's stock price slightly declined, but this fluctuation did not hinder its rapid pace. It took only 66 trading days for its market value to exceed $2 trillion.

Based on the closing price on June 6th, NVIDIA only needs to rise another 6% to surpass Microsoft and become the world's highest market value company. Some investment banks predict that NVIDIA's ascent to the top is almost certain, and it is expected to be achieved as soon as the stock split is completed next week.

This seems like a metaphor, indicating that the era of AI has definitely arrived.

Strong market demand and investor support have lifted NVIDIA to the throne. However, the market is increasingly worried about whether the next "Cisco moment" will come. Once NVIDIA shows signs of decline in sales and profit margins, the celebration will turn back into silence.

At the top, Huang Renxun must harness the power of AI to make NVIDIA the core infrastructure of the artificial intelligence era, firmly binding various tech giants to himself in order to firmly hold onto power in this new era.

Soaring

NVIDIA has achieved an unprecedented myth in the U.S. stock market. Previously, it took Apple over two years to go from $2 trillion to $3 trillion, while NVIDIA created a miracle in just three months.

At the same time, CEO Huang Renxun's wealth has also skyrocketed, with a net worth exceeding $100 billion, making him one of the "super-rich" with assets exceeding $100 billion on Forbes' global rich list.

The soaring market value of NVIDIA lies in the gradual dissipation of the capital market's divergence towards it. Outstanding performance, amazing products, and the excitement of investors trading due to the stock split have made Wall Street believe that it can still break through the "ceiling".

In terms of performance fundamentals, NVIDIA's total revenue in the first quarter reached $26 billion, more than three times that of a year ago. NVIDIA, selling "AI weapons", is steadily moving towards its peak, and this number is expected to continue to climb to $28 billion in the second quarter.

On June 2nd, Huang Renxun, who appeared at the 2024 Taipei International Computer Show, became the most eye-catching "star" of the event. He announced that the Blackwell, which had been in production for less than 3 months, was already in production. More importantly, NVIDIA will release a new generation of products every year, with a three-year upgrade plan emerging: Blackwell Ultra in 2025, new architecture Rubin iteration in 2026, and Rubin Ultra listed in 2027 At the same time, to continue strengthening its moat, NVIDIA has long turned itself into an "AI factory." In the presentation, the "Leather Jacket Master" showcased a powerful ecosystem, based on GPUs and CUDA, using various chips and devices, users can continuously link computing power to form a huge data center, becoming the "power plant" of the AI era.

Whether it is the ultimate pursuit of software and hardware ecosystems or the annual upgrade rhythm, NVIDIA's "strategy" is to firmly bind a group of tech giants to its own "AI ecosystem," leaving no room for opponents to catch their breath.

Wall Street is quite excited about this, with Bank of America raising NVIDIA's target stock price to $1500. Analysts believe that AI is currently developing rapidly, with many top tech and cloud computing companies still in the experimental stage of training large language models, so the peak of AI in the chip industry has not yet arrived.

It is estimated that the annual investment in updating computing power in global data centers needs to be around $250-500 billion, and currently only 20-30% has been completed, with another 3-5 years needed to complete.

Institutions generally believe that NVIDIA surpassing Microsoft to become the world's largest market cap is only a matter of time. However, Huang Renxun hopes to add more fuel to the fire.

Not long ago, NVIDIA announced a "1-for-10" stock split on June 7th, allowing more retail investors to board their fast train and provide momentum for the stock price to continue to rise. The stock split will also make it easier for NVIDIA to join the Dow Jones Index and consolidate its position in the U.S. capital market.

This series of layouts are all-round preparations made by Huang Renxun for NVIDIA to become the "top player" in the global capital market.

Challenge

From the current perspective, NVIDIA is undoubtedly a gem sought after by the masses. Over the past 5 years, NVIDIA's stock price has risen nearly 30 times, and this year it has surged over 150%.

The story of the elephant dancing is not uncommon in the U.S. stock market, but the speed at which it is happening still leaves investors amazed. From 1992, when the idea of creating NVIDIA burst forth from a chain restaurant in Silicon Valley with other founders, to now becoming a giant, Huang Renxun's NVIDIA has also teetered on the brink of bankruptcy several times.

Investment banks are also concerned that the bubble of this AI frenzy may be approaching a critical point, and NVIDIA may experience a "Cisco moment" again.

As the "arms dealer" of the last Internet revolution, Cisco's market value evaporated by about 90% in the years following the bursting of the Internet bubble and has never returned to its peak since.

The market's direction changes in the blink of an eye, and the semiconductor industry is also highly cyclical. Whether it's cryptocurrency or the gaming craze during the pandemic, NVIDIA has experienced significant fluctuations in revenue and profits due to surging short-term demand. The bursting of the cryptocurrency bubble in 2021 once led to NVIDIA being abandoned by the market, with the stock price dropping from a historical high of $835 to less than $200 in just half a month.

Currently, NVIDIA is still in a state of continuously breaking revenue and profit margin records, with products in high demand and "selling well."

However, the existence of cycles indicates that if NVIDIA wants to continue to refresh its history, both sales and profit margins must maintain a strong momentum, while also building up its moat. On the way to the top, NVIDIA is facing a series of challenges from industry trends, attacks from competitors, resource bottlenecks, and more At the current juncture, NVIDIA occupies over 70% of the market share for AI training chips, with a dominant 92% market share in data center GPUs. This large AI market share is hard to resist for competitors.

AMD CEO Lisa Su bluntly pointed out that NVIDIA's dominant position seems to go against the industry's will. She stated, "I don't believe in the moat theory of NVIDIA. It is important whether customers have a choice. They actually hope to have more opportunities for hardware experimentation."

Soon after, in early June, AMD unveiled the new MI350X AI accelerator chip. Compared to NVIDIA's most powerful AI chip, the H200, the memory, bandwidth, and computing performance have all doubled. Similar to NVIDIA, AMD also aims to release a new product every year. Intel also revealed that its Gaudi 3AI chip will be priced much lower than its competitors, with plans for a large-scale launch in the third quarter of this year. According to official positioning, this chip's performance is 1.5 times that of NVIDIA's H100, equivalent to the H200.

Moreover, not only that, but major customers of NVIDIA such as Microsoft, Amazon, Google, Meta, who have no better choice but to rely on NVIDIA, have also started to "make their own weapons" by designing AI chips and commissioning manufacturers for production, attempting to reduce their dependence on the NVIDIA leader.

More importantly, after the fierce battle of large models, both tech giants and AI startups have entered a period of calm in finding scenarios for large models to land.

In March, Sequoia Capital estimated that AI startups sent $50 billion to NVIDIA for training large language models, but only received a meager $3 billion in revenue from the market.

In other words, in this grand "arms race," players must rethink the reality of achieving profitability. If commercialization is delayed, NVIDIA may also find it difficult to escape the cycle of the industry.

Standing at the peak of the industry and being swept along by trends, Jensen Huang has no way out. He must wage a defensive battle for AI infrastructure, waiting for the industry's average return and the ebb of high profits, still able to stand out among the crowd