The AI chip giant, known as the most important stock on earth, according to Goldman Sachs,$NVIDIA (NVDA.US)$Despite the pressure from heavy selling by company insiders and the sale of over $700 million in stocks by Nvidia insiders this year, demand for the company's AI chips has continued to skyrocket globally. Although the stock faced intense selling pressure after the May performance announcement, Nvidia's stock price has continued to climb on record highs and the bullish trend may be far from over.

According to statistics compiled by Washington Service, Nvidia executives and directors have sold approximately 770,000 shares of Nvidia stock, worth more than $700 million, so far this year, not including the impact of the stock split on June 10, where stocks were split in a 1-to-10 ratio. According to the latest data from Washington Service, this is the largest reduction in holdings in terms of half-yearly basis since the first six months of 2023, when about 848,000 shares were sold. This year's dollar value of stock sales is by far more substantial than the previous rounds of internal reductions due to the stock's amazing rise in value since 2024, which has reached an astonishing 164%. As various global businesses continue to compete in an arms race for increasing the overall computing power of AI systems, AI chips are essential in driving high-efficiency generated AI tools such as data center servers' AI accelerators chips (i.e., AI chips represented by Nvidia's AI GPU), thus driving up the strong demand for these chips. Undoubtedly, Nvidia dominates the server AI chip market with a share of up to 90%.

"Storm of insider selling" sweeps Nvidia

In late May, Nvidia, known as the AI chip giant, once again announced unparalleled performance that shocked global investors, dispelling any concerns for a slowdown in spending on AI-related businesses. Nvidia has single-handedly reinforced technology investors' belief in AI, driving technology stocks in the US stock market to continue to rise recently and also boosting Nvidia's stock price to start a new round of crazy price increase.

Nvidia's total Q1 revenue in 2024 increased by 262% to US $26 billion, with total revenue reaching a historic high, and the YoY growth rate of total revenue exceeding 200% for three consecutive quarters. Under the impetus of strong demand for H100/H200 GPUs, Nvidia's Q1 data center revenue reached a historic high of $22.6 billion, up 427% YoY.

"Storm of insider selling" sweeps Nvidia

Mark Lehmann, CEO of Citizens JMP Securities, believes that the large-scale insider selling is indeed worth paying attention to, but considering that some executives are compensated in the form of shares, and there seems to be no sign that demand from global companies for the company's high-performance AI GPU products is showing any signs of slowing down, this may not necessarily cause market panic.

"Whenever you see this kind of wealth effect and this scale of market cap being created, I'm always looking for the people who came and went, and I didn't see executives who helped us achieve this scale making massive sell-offs of holdings." Lehmann said in an interview. "That would have worried me more."

This chip giant is currently the world's third most valuable listed company, with a market capitalization of approximately $32.2 trillion, second only to US technology giants Microsoft (MSFT.US) and Apple (AAPL.US). The gap between the three companies is very small, and Apple and Microsoft have recently taken turns leading in total market capitalization with each other, enjoying the throne of "the world's highest market value listed company".

Statistics show that more than one-third of Nvidia insiders who chose to sell stocks this year did so after the company released another unparalleled earnings report on May 22. Nvidia also announced that it would continue to divide stocks shortly thereafter, these positive factors have led to a new round of crazy increase in the company's stock price. During this period, the biggest contributor to the large selling pressure were the two Nvidia board members, Mark Stevens and Tanchi Cox.

On Monday in US Eastern time, Jensen Huang, the CEO of Nvidia, known as the "father of AI", reported that he had sold approximately $31 million worth of stocks according to a pre-arranged trading plan. A Nvidia representative declined to comment.

Although there are many insider sellers, there has been a lack of internal buying power. According to statistics compiled by Washington Service, Nvidia has not had any internal purchases of stocks since its CFO, Colette Kress, bought a large amount of stocks in December 2020, except for exercising important options.

As Nvidia executives and directors sell off their shares in droves, the stock prices of US technology stocks closely related to AI also show signs of overheating. On Monday, large tech giants including Nvidia, Microsoft, Google, and Ethernet switching chips and ASIC chip giant Broadcom (AVGO.US) all rose further, pushing the Nasdaq 100 index to set further new highs.

However, the 14-day relative strength index (RSI) of the Nasdaq 100 index, which covers many popular tech stocks, is currently at the 'highest overbought level' since 2018, meaning that the index could experience a sharp pullback at any time.

From a long-term perspective, the rise of 'the most important stock on earth' may be far from over.

Beth Kindig, a technology industry analyst from the well-known investment firm I/O Fund, recently released a research report predicting that Nvidia's stock price will soar by about 258% from current levels by 2030, and its market cap is expected to reach $10 trillion by then (Nvidia's current market cap is about $297 billion). The main logic is that Nvidia's "CUDA+AI GPU" ecosystem brings an incredibly powerful moat, and its next-generation AI GPU based on the Blackwell architecture is expected to bring huge revenue contributions.

Nvidia has been deeply rooted in the global high-performance computing field for many years, especially its CUDA computing platform, which is recommended by software and hardware for high-performance computing applications such as AI training/inference. The CUDA computing platform is an exclusive parallel computing acceleration platform and programming assistant software developed by Nvidia, which allows software developers and software engineers to use Nvidia GPUs to accelerate parallel general-purpose computing (only supports Nvidia GPUs and cannot be compatible with mainstream GPUs such as AMD and Intel).

CUDA can be said to be the platform on which generation-based AI applications such as ChatGPT depend heavily. Its importance is on the same level as the hardware system and is crucial for the development and deployment of large-scale AI models. With extremely high technical maturity, absolute performance optimization advantages, and extensive ecosystem support, CUDA has become the most commonly used and widely popularized collaborative platform in AI research and commercial deployment.

Nvidia's currently hottest AI chip, the H100/H200 GPU accelerator, is based on Nvidia's breakthrough Hopper GPU architecture, providing significantly more powerful computational power than its predecessor, especially in floating-point operations, tensor core performance, and AI-specific acceleration. Even more significant is that the AI GPU based on the Blackwell architecture has far higher performance than the Hopper architecture. In particular, the GB200 inference performance of the Blackwell architecture on the GPT-3 LLM benchmark with 175 billion parameters is 7 times that of the H100 system, and provides four times the training speed of the H100 system.

Regarding when the new Blackwell architecture AI GPU will be put into various data centers, Nvidia CEO Jensen Huang said that the brand-new Blackwell architecture AI GPU product will be shipped in the second quarter of this year, increased in the third quarter, and officially put into data centers in the fourth quarter. It is expected that the 'Blackwell architecture chip revenue will grow significantly' this year. In the Blackwell architecture AI GPU press release issued by Nvidia in March, Tesla CEO Musk publicly stated that Nvidia's AI hardware was the 'best AI hardware'.

Analyst Jin Di predicts that by the end of Nvidia's fiscal year 2026, revenue from the Blackwell architecture AI GPU will significantly exceed that of its predecessor H100 architecture, with Blackwell architecture expected to drive Nvidia's datacenter revenue to as high as $200 billion.

At the top investment bank on Wall Street, the bullish voice on Nvidia's stock price is constant. Bank of America pointed out that although Nvidia's stock price has repeatedly hit historical highs recently, there is still huge upside potential because the chip company is expected to continue to dominate the AI chip market in the next few years. In a report released on June 5 (Wednesday), the bank reiterated its 'buy' rating on Nvidia's stock and stated that the company led by Huang is still the top choice in the IT industry. Bank of America's strategist set a target price of up to $150 within 12 months for Nvidia (Nvidia's stock price closed at $130.98 on Monday).

Susquehanna raised its target stock price for Nvidia from $145 to $160 in a recent report. The institution stated in the report: 'Although our valuation is far higher than the average of peer companies at about 28.5 times, we believe it is reasonable because Nvidia can seize significant opportunities in the booming AI terminal market.'

Rosenblatt, a well-known institution on Wall Street, raised Nvidia's target stock price from $140 to $200, maintaining a 'buy' rating. This institution's price of up to $200 is the highest target stock price on Wall Street within the next 12 months. In addition, Barclays Bank raised Nvidia's target stock price from $120 to $145, and Argus Research raised Nvidia's target stock price from $110 to $150.

Edited by Jeffrey