Nvidia has become the most valuable stock in history. Huang (Jensen Huang) now has a net worth of $138.8 billion, and even the newly - hired researchers have caught this wave.

What's the concept? According to LSEG data, this figure exceeds the combined total market value of the stock markets in Canada and Mexico and also surpasses the total market value of all listed companies in the UK.

Previously, the holder of the all - time high record was Apple. On December 26th last year, Apple set a record closing value of $3.915 trillion.

At the same time when Nvidia set a record, among the listed companies on Wall Street, Microsoft's total stock market value reached $3.7 trillion, ranking second; Apple's total market value was $3.19 trillion, ranking third.

The sky - rocketing market value of Nvidia is enough to show the current popularity of the AI track. In 2023, Nvidia's market value first exceeded $1 trillion. In the past two years when AI has been booming, this figure has doubled rapidly.

According to Forbes' calculation, at this moment Jensen Huang's personal net worth is $138.8 billion, an increase of $1.8 billion from before, a growth rate of 1.31%, and he currently ranks 10th on the global rich list.

A market value of $3.92 trillion and a stock price of $160.98

According to Reuters, Nvidia's stock price soared 2.4% in early trading, reaching $160.98 per share after the increase. As a result, its total market value reached $3.92 trillion, making it the "company with the highest market value in history".

However, subsequently, the increase in its stock price narrowed to 1.5%, with a per - share quote of $159.60, corresponding to a market value of $3.89 trillion. Although it failed to maintain the leading position continuously, it still closely approached Apple's all - time peak with a narrow margin.

△

This market - value frenzy stems from Wall Street's unprecedentedly high optimism about AI.

The core driving force is the technological breakthrough of Nvidia's latest - generation AI chips in training ultra - large - scale models.

Currently, tech giants such as Microsoft, Apple, Amazon, Meta, Google, and Tesla are engaged in an AI data - center arms race, and their demand for Nvidia's high - end processors has skyrocketed.

As Joe Saluzzi, co - manager of Themis Trading, said: "The leap from a trillion - dollar market value to approaching four trillion confirms that AI spending is setting off a super - boom, and the entire market is chasing this technological revolution."

Just a few hours ago, hardware media tom’shardware reported that Nvidia's latest - generation flagship AI super - computing system completed its first commercial deployment globally.

Dell announced that it has delivered the industry's first systems built on the Nvidia GB300 NVL72 platform to CoreWeave, a leading AI cloud - service provider. CoreWeave said that it has deployed clusters based on the Blackwell Ultra super - chip through its data - center hosting provider, Switch.

A single GB300 NVL72 rack is like a miniature computing - power factory:

72 Nvidia Blackwell Ultra super GPUs and 36 Arm - based 72 - core Grace CPUs form a heterogeneous computing matrix, paired with 20TB of HBM3E high - speed video memory and 40TB of system memory, unleashing 1.1 ExaFLOPS of FP4 inference computing power and 0.36 ExaFLOPS of FP8 training performance.

It has a 50% performance improvement compared to the previous - generation GB200.

Behind CoreWeave's aggressive procurement strategy is the harsh reality of the AI computing - power market: when the cost of training a top - tier large model exceeds ten million dollars per session, every 1% improvement in hardware performance means real savings.

In addition, the soaring stock price of Nvidia may also be related to the peak shipment period of the GB200.

The shipment of the GB200, originally expected to start at the end of 2024 or the beginning of 2025, was postponed due to supply - chain challenges. Previous supply - chain information indicated that the peak shipment period was expected to be postponed to between the second and third quarters of this year, and it has indeed arrived.

When the AI strategies of giants like Microsoft and Google are built on its hardware foundation, the expansion of this chipmaker's market value is actually a collective vote by the entire technology industry for the concept that "computing power is productivity".

Currently, Nvidia accounts for 7% of the S&P 500 (.SPX) index. Together with Microsoft, Apple, Amazon, and Alphabet, they account for 28% of the index, which means that the wealth of tens of millions of investors who rely on index funds for retirement is deeply tied to the future of AI technology.

However, the market is not entirely optimistic. Kim Forrest, the chief investment officer of Bokeh Capital, said: "AI is indeed a revolutionary tool, but the actual value of current large - language models may not match the market hype."

Meanwhile, the launch of the domestic large - language model DeepSeek in January this year, which outperforms many top - tier models in terms of cost - effectiveness, has also led to market speculation that companies may reduce their spending on high - end processors, triggering a global stock - market sell - off.

Now, this chip - driven market - value frenzy is testing the market's real ability to digest technological changes.

While the AI chip industry is booming, Nvidia has not stopped poaching talents.

Jensen Huang has been poaching Tsinghua students wave after wave

Recently, following the acquisition of LeptonAI by Yangqing Jia, the Tsinghua - affiliated AI company Nexusflow was acquired by Nvidia.

Then, Jensen Huang rushed to poach talents himself, including:

Jiantao Jiao, the winner of the Undergraduate Special Award at Tsinghua University in 2011. Former CEO of Nexusflow, associate professor at UC Berkeley, and graduated with a doctorate from Stanford. He joined Nvidia as a research director and Distinguished Scientist.

Banghua Zhu, also an undergraduate from Tsinghua. Former co - founder of Nexusflow, will join the University of Washington as an assistant professor and graduated from UC Berkeley. He joined Nvidia as a Principal Research Scientist.

Jian Zhang, an alumnus of Tsinghua. Former CTO of Nexusflow, became the director of applied research at Nvidia.

It's worth mentioning that Nexusflow's connection with Nvidia is related to Kurt Keutzer, another co - founder and a professor at UC Berkeley. He participated in the Efficient AI project, a collaboration between Nvidia and the academic community in 2025.

The Efficient AI project was originally jointly initiated by Nvidia and MIT, and the person in charge is Song Han, an associate professor at MIT EECS.

Later, OmniML, founded by Song Han and focused on optimizing machine learning on edge devices, was also acquired by Nvidia in July 2023. Song Han joined Nvidia as a Distinguished Scientist.

Di Wu, the CEO of OmniML, and Huizi Mao, the CTO, also joined Nvidia. Di Wu became the director of deep - learning algorithms and software, and Huizi Mao became the technical supervisor and senior engineer of the deep - learning algorithms and software team.

Song Han, Di Wu, and Huizi Mao all completed their undergraduate studies at Tsinghua.

Although Yangqing Jia has been well - known for a long time, he graduated from the Department of Automation at Tsinghua University.

Calculated by age, these people can be divided into three groups. Yangqing Jia entered Tsinghua in 2002, Di Wu, Song Han, and Jiantao Jiao entered between 2007 - 2008, and Banghua Zhu and Huizi Mao entered between 2012 - 2014.

These waves of Tsinghua students started AI - related businesses one after another, and now they have reunited at Nvidia.

Leaving other things aside, they really caught the best time for stock - option salaries.

Reference links:

[1]https://www.reuters.com/business/nvidia-set-become-worlds-most-valuable-company-history-2025-07-03/

[2]https://finance.yahoo.com/news/3-92-trillion-shock-nvidia-214445904.html

[3]https://www.tomshardware.com/tech-industry/artificial-intelligence/nvidias-newest-top-tier-ai-supercomputers-deployed-for-the-first-time-grace-blackwell-ultra-superchip-systems-deployed-at-coreweave

[4]https://www.forbes.com/profile/jensen-huang/

This article is from the WeChat official account "QbitAI", written by Xifeng, and published by 36Kr with permission.