36 people reporting to Jensen Huang at NVIDIA

In the market value giant NVIDIA, who can report directly to CEO Jensen Huang?

36 people.

This is the latest disclosed number.

These 36 individuals are roughly divided into seven functional areas—

Strategy, Hardware, Software, AI, Public Relations, Networking, and one executive assistant to Huang.

Huang's New Layout

Next, let's take a look at what pieces Huang has laid out in this year's personnel arrangement, aside from the highest strategic execution level, and what signals he has released.

First, hardware remains the cornerstone of NVIDIA.

Among Huang's direct subordinates, there are 9 responsible for hardware-related businesses—including GPU, telecommunications, DGX systems, etc.—accounting for one-third of the total.

Whether it's the AI bubble or the CUDA ecosystem, NVIDIA's fundamental color is still hardware.

It is worth noting that AI, embodied intelligence, and autonomous driving technologies are gradually becoming the "second pillar" in Huang's business landscape.

In this area, Huang has deployed seven key figures, including Eddie Wu.

From the configuration of personnel alone, it can be seen that NVIDIA's reach is rapidly extending into those new territories that have not yet been fully explored, which Huang refers to as the "zero billion dollar market."

This journey is destined to be a protracted battle, and if lucky, it may even become Huang's final campaign in his career.

Surprisingly, among Huang's direct executives, there are actually three responsible for public relations.

You might not have a concept of this number; for example, Musk doesn't even have a public relations manager.

The reason lies in the completely different ecological niches the two occupy.

NVIDIA provides computing power, which is upstream of the entire industry, and is closely related to practitioners in various fields. They need to be responsible for global enterprises, research institutions, government projects, and investors.

This means that NVIDIA, in addition to making products, also has to spend a lot of effort to clarify the intricate industrial chain—

They need to appease Wall Street while maintaining the developer ecosystem; serve large clients while considering upstream and downstream partners; and even, during special periods, coordinate policies among countries.

Therefore, Huang needs a systematic external communication mechanism.

However, the situation of someone like Musk, who completely does not set up a public relations team, is relatively unique.

After all, Musk has long operated himself as the world's largest "super influencer," and he himself is the strongest PR and GR.

Old Friends and New Faces

After reviewing Huang's overall layout, let's take a look at the key generals under his command First, let's take a look at three old friends— Jonah Alben, Dwight Diercks, Bill Dally.

These individuals have been working alongside Jensen Huang for a long time, and those who have read Huang's biography should often see their names.

Jonah Alben

The first to appear is the leader who has built half of NVIDIA's empire, referred to by Jensen Huang as the "soul of GPU architecture"— Jonah Alben.

This year marks Alben's twenty-eighth year at NVIDIA.

Since 2008, Alben has served as Senior Vice President of GPU Engineering at NVIDIA, fully responsible for the design and development of GPU architecture.

Prior to this, he served as Vice President of GPU Engineering for four years, playing a key role in NVIDIA's transition from discrete graphics cards to the AI era.

Alben joined NVIDIA in 1997, the year the company launched its first generation of RIVA series GPUs. His initial position was as an ASIC design engineer, participating in the early graphics chips of NVIDIA.

Before joining NVIDIA, he also worked as an ASIC engineer at Silicon Graphics.

Alben graduated from Stanford University, where he obtained a Bachelor's degree in Computer Science and Engineering and a Master's degree in Electrical Engineering.

Currently, Alben holds a total of 34 patents and manages a GPU engineering team of over a thousand people.

Dwight Diercks

Next is another cornerstone supporting NVIDIA's business empire—Jensen Huang's technical right-hand man, in charge of software, Dwight Diercks.

Diercks has been working with Jensen Huang for 31 years and is likely the most senior executive among this group of direct subordinates, aside from the co-founders.

As early as 1994, Diercks began his career at NVIDIA as a Senior Software Engineer.

At that time, NVIDIA had just been established, and he was the 22nd employee to join. Over these thirty-plus years, he has accompanied Huang through all the ups and downs.

Many of the NVIDIA stories circulating in the outside world come from his accounts.

In 1999, Diercks was promoted to Vice President, officially overseeing the software development system. By 2017, NVIDIA had just over ten thousand employees, but more than three thousand software engineers reported directly to him Eight years have passed, and NVIDIA's scale has more than tripled. It's hard to imagine how large the team under Diercks is now.

Currently, he serves as NVIDIA's Executive Vice President of Software Engineering, fully responsible for developing core system software and platform layer support for all product lines, including PC and workstation graphics cards, deep learning accelerators, autonomous driving platforms, AI frameworks, cloud computing, and gaming devices.

Before joining NVIDIA, Diercks worked as a systems software engineer at Pellucid Inc. and also worked at Compaq, engaging in multimedia software development.

Diercks graduated from Milwaukee School of Engineering in 1990 with a Bachelor's degree in Computer Science and Engineering and received an honorary Doctor of Engineering degree from the same school in 2014.

Bill Dally

Now let's take a look at Bill Dally.

Dally is NVIDIA's Chief Scientist.

Before joining NVIDIA, he taught at Stanford University for 12 years, serving as the chair of the Computer Science Department and was recognized as an authority in the field of parallel computing.

At that time, he often praised NVIDIA.

In 2003, Jensen Huang personally visited his university office and extended an olive branch for a part-time consulting position.

Six years later, under NVIDIA's persistent persuasion, Dally, who was supposed to dedicate his life to academia, ultimately agreed to become a full-time employee at NVIDIA.

Since then, Dally has become the soul of NVIDIA's R&D system. He not only promoted the evolution of GPUs from graphics processors to general-purpose parallel computing platforms but also laid the foundation for the birth of AI hardware architecture.

Not wanting to endure history classes, Dally dropped out of high school. During his time as a car mechanic, he managed to enter college based on his exam scores.

Although he did not obtain a high school diploma, he earned a Bachelor's degree from Virginia Tech, a Master's degree from Stanford University, and a Ph.D. from the California Institute of Technology.

In his early 30s, Dally was already a tenured professor at MIT.

However, this scholar's life was not limited to the study.

In 1992, he flew a single-engine plane to New York but encountered an oil leak mid-flight, forcing an emergency landing in the Long Island Sound.

The violent impact when the plane hit the water fractured his nasal bone and left him dizzy, nearly sinking with the plane.

Fortunately, a passing sailboat rescued him in time.

Two days later, he returned to the lab as if nothing had happened.

After looking at these three veterans, let's take a look at a fresh blood from NVIDIA, who is also the only Chinese among Eric Wu's direct subordinates.

Eddie Wu

Eddie Wu is currently the Vice President of Automotive Business at NVIDIA, fully responsible for the strategic planning, product layout, and engineering execution of the automotive business.

Eddie Wu graduated from the Department of Electrical Engineering at Tsinghua University and later pursued a master's and doctoral degree in electrical engineering at the University of Illinois at Urbana-Champaign, holding over 250 authorized U.S. patents.

From 2006 to 2018, Eddie Wu worked at Qualcomm, where he was responsible for several core R&D projects, making significant contributions to Qualcomm in the fields of autonomous driving, precise positioning, and communication technology.

His connection with XPeng began in 2015.

At that time, Qualcomm was focusing on the autonomous driving industry and planned to acquire NXP Semiconductors. It was during this phase that Eddie Wu was entrusted with the important role of head of Qualcomm's autonomous driving business.

Unfortunately, this acquisition ultimately fell through due to "antitrust" issues, but it also allowed Eddie Wu to meet an "important man" in his life—He Xiaopeng.

On one side was Qualcomm's failed acquisition; on the other side was He Xiaopeng, who was eager to build an autonomous driving team and was in search of talent.

Thus, the two sides hit it off. Eddie Wu became the Vice President of the Autonomous Driving Center at XPeng, reporting directly to He Xiaopeng and leading a technical team of over a thousand people.

During his five years at XPeng, he fully led the strategic layout, product planning, and technical implementation of autonomous driving, promoting the launch of core projects such as NGP and XNGP, paving the way for mass production of autonomous driving by Chinese brands.

The XPeng G6 can be regarded as the culmination of Eddie Wu's autonomous driving career. The model surpassed 40,000 orders within just one month of its launch, with the "Max version," equipped with the high-end intelligent driving system, accounting for as much as 70%.

However, just as XPeng was gradually getting on track, another "important man" appeared in Eddie Wu's life—Jensen Huang.

It is reported that during his tenure at XPeng, Eddie Wu had many interactions with NVIDIA due to project collaborations.

At that time, NVIDIA had firmly established itself as the global leader in the high-performance computing SoC market for autonomous driving, with a market share of 82.5%. However, despite its technological lead, its automotive business still accounted for a very limited share of the company's total revenue.

At this critical juncture, NVIDIA was determined to accelerate its layout in the autonomous driving business, and Eddie Wu, with experience in both algorithms and vehicle manufacturers, who understood underlying computing power and was familiar with automakers' needs, was undoubtedly the ideal candidate in Jensen Huang's eyes.

Upon leaving, Eddie Wu humorously remarked on Weibo: Although I am at NVIDIA, I am still a "Peng" friend.

Tomorrow is my first day at NVIDIA. Thanks to XPeng for personally sending me to Eric Wu. According to Eric Wu, I will still be working for XPeng in the future, just without having to pay my salary

He also stated that He Xiaopeng and Jensen Huang are the two most important men in his career.

After moving to the United States to join NVIDIA, Wu Xinzhao proposed the concept that "all movable things will eventually move towards automation," and brought the engineering experience he accumulated at Qualcomm and XPeng into NVIDIA's autonomous driving system. He led the system optimization of AI models for perception and decision-making, driving NVIDIA to complete the upgrade from hardware-driven to fully self-developed in the autonomous driving technology stack.

At the same time, NVIDIA's autonomous driving team in China is also rapidly expanding, continuously improving the algorithm and data closed-loop system, and significantly enhancing platform capabilities.

These contributions are quantifiable—during the fiscal year 2024 to 2025, NVIDIA's automotive business revenue soared from $281 million to $567 million, nearly doubling.

In contrast, before Wu Xinzhao joined, NVIDIA's automotive business revenue declined by 4% throughout 2023.

One must say, Jensen Huang has a keen eye for talent.

Structural Reorganization During the Expansion Period

Having 36 direct subordinates is quite a large team for the CEO of a technology giant with a market value of $4 trillion.

Such a terrifying management intensity, even for the workaholic Elon Musk, known for his "hands-on" approach, seems slightly inferior.

As of August, Musk had only 19 direct subordinates at Tesla; and in his newly founded xAI company, only 5 executives report directly to him.

In fact, 36 is not the upper limit of the number of direct reports for Jensen Huang.

Previously, in an interview at Stanford University in March 2024, Jensen Huang stated that he had 55 direct subordinates.

It is well known that Huang has always been a staunch advocate of a flat organizational structure.

In 2023, at the DealBook Summit of The New York Times, Huang stated that the shorter the decision-making chain, the faster the flow of information.

The more direct reports a CEO has, the fewer levels there are in the company. This allows us to maintain smooth information flow and ensure that everyone has access to the information they need.

To manage such a large team, Huang is quite persistent about the concept of "information transparency."

Unless someone specifically requests it, he usually does not arrange one-on-one meetings with direct subordinates; instead, everything is discussed together so that information can spread freely.

Musk shares the same view on this.

If cross-departmental communication is to be achieved, people should not have to go through the entire chain of command to relay messages, but should be allowed to communicate directly to achieve goals in the shortest path.

Their preference for a flat management model is not merely a mutual admiration between heroes; to a large extent, it is a trend in the entire technology industry.

In the high-tech field, the speed of product iteration is astonishing, which is vividly reflected in the AI industry, where the frequency of model updates is almost fast enough to be measured in "weeks." Against this backdrop, a flat organizational structure can effectively weaken hierarchical approvals, allowing the CEO or core leaders to make decisions more quickly.

NVIDIA's success with the Blackwell architecture and DGX systems is attributed to this "parallel advancement and rapid decision-making" research and development model.

Moreover, a flat company structure also provides the CEO with more opportunities to directly engage in the company's actual business.

Around 2020, Jensen Huang required every employee to submit a list each week, outlining their five most important tasks.

Since then, every Friday, he would receive about 20,000 emails and often browse through them randomly late at night.

Correspondingly, he also writes hundreds of emails each day.

Maintaining communication with frontline personnel over the years has gradually honed Huang's exceptionally keen intuition for the company's "pulse," enabling him to capture subtle changes in technology, market, and supply chain in real time.

Regarding the Transformer architecture, Huang learned about it through these tens of thousands of emails.

However, this extremely flat-seeking CEO has seen his direct reports shrink by nearly 40% in less than two years.

This inevitably raises curiosity—what exactly has happened at NVIDIA?

We currently have no way of knowing the internal news, but starting from the company's financial reports at several key points, we might glimpse some clues.

The most intuitive aspect is, of course, the explosive growth of NVIDIA's business.

In fiscal year 2024, NVIDIA achieved a net profit of approximately $29.5 billion, soaring nearly 600% year-on-year; just three months into fiscal year 2025, the net profit has already climbed to $14.8 billion, a staggering increase of 628% year-on-year.

The soaring profits have also led to a dramatic expansion of NVIDIA's organizational scale.

As of early 2024, NVIDIA had a total of 29,600 employees; just one year later, this number grew to 36,000, an increase of 21.62% year-on-year.

This marks the largest expansion of the company's scale in the past 16 years, except for 2021.

In this context, the limitations of the original flat model have begun to surface.

The flat system is indeed efficient: short decision chains, strong execution, and top management can directly grasp frontline information.

However, when the organizational scale rapidly expands and personnel sharply increases, the volume of new information generated daily also grows exponentially, leading to huge information noise.

In this situation, the cost of cross-departmental collaboration suddenly rises, and if management insists on personally handling every aspect of the company's business, it can easily lead to burnout.

In response to this issue, Musk's choice was—direct layoffs.

In 2022, after acquiring Twitter, the number of Twitter employees plummeted from over 7,000 to less than 2,000 in just a few months, with about 80% of employees being asked to pack up and go home In contrast, Jensen Huang is portrayed as a "good boss." Although he enjoys publicly berating his subordinates, he rarely fires anyone.

In 2009, due to a mistake by a chip architect, the company mistakenly connected a chip to the solder "bumps" under the circuit board, resulting in a large number of product failures.

To resolve this issue, Huang had to allocate $200 million for customer refunds, causing NVIDIA's annual profit to drop to zero and its market value to plummet by nearly 90%.

Despite this, that architect ultimately kept his job.

As one NVIDIA executive said:

He will scold you, yell at you, and even insult you, but no matter what, he will never fire you.

Even in a poor business environment, when he had to close a department, he would do his best to reassign employees to other needed positions.

However, with the company's rapid expansion and a desire to avoid layoffs, the only choice was to abandon the flat structure, dispatch more senior executives, and move towards a vertical organizational structure.

Although Huang himself has never publicly acknowledged this, the noticeable reduction in the number of direct subordinates likely signals a significant turning point in his management model—

NVIDIA has gradually slowed down from being a fast-running follower to becoming a more stable and mature scaled enterprise.

Iron-fisted pressure, starkly different from Silicon Valley's style

While Huang does not actively fire people unless absolutely necessary, the number of employees who choose to leave on their own is likely not small.

Internally, Huang has always advocated a rigorous high-pressure culture, constantly emphasizing that the company is in a critical moment of life and death.

In 1999, NVIDIA, which had been established for less than six years, went public with a valuation of $600 million, and the stock price subsequently soared, making Huang a millionaire overnight.

However, the employees did not receive a celebratory banquet filled with toasts, nor even a congratulatory message from the boss.

On the second day after the IPO, Huang sent an internal letter to the entire company:

We must complete the task at all costs; time is of the essence. We must do everything we can to win this victory.

Remember, our priorities are first, second, and third... we must complete the task.

Huang's philosophy is clearly embodied in NVIDIA's office building.

In Silicon Valley, NVIDIA's office building seems out of place: there are no gyms, no climbing walls, no dog parks, and no disc golf courses.

As for the reason... Huang feels it's unnecessary.

Everyone is here to work.

With such a workaholic boss, long hours naturally become the norm, which many new employees find hard to bear.

According to one employee's recollection, every day is filled with endless deadlines.

The result is an almost endless stream of deadlines, always feeling like there's not enough time If compared to the previous company, it is even more tear-jerking.

At 3dfx, our motto is "Work hard, play hard," but at NVIDIA... it's just "Work hard."

What is Jensen Huang's way of thinking? Just look at how he evaluates "second place" to find out—

Second place is the first loser.

This article is sourced from: Quantum Bit, original title: "The 36 People Reporting to Jensen Huang at NVIDIA"

Risk warning and disclaimer

The market has risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investing based on this is at your own risk