In November, NVIDIA released its fiscal 2025 Q3 financial report, revealing revenues of $35.1 billion and a GAAP net profit of $19.309 billion, marking a year-on-year increase of 109% and surpassing market expectations.

NVIDIA has long been a pioneer in accelerated computing, driving the PC gaming market and revolutionizing computer graphics with its GPUs. Its CUDA platform and programming model have significantly impacted AI technology, ushering in the modern AI era.

NVIDIA's success stems not only from its technological prowess but also from Jen-Hsun Huang's innovative management strategies. These strategies have fueled the creativity and productivity of NVIDIA employees, ensuring a steady stream of cutting-edge products and maintaining the company's leadership in global technology.

How did NVIDIA achieve this success and sustain its leadership? 'The NVIDIA Way: Jen-Hsun Huang and His Tech Empire,' published by CITIC Press, delves into Huang's unique insights and methods in technological innovation, market competition, business management, talent recruitment, and equity incentives.

I. Two Key Market Strategies

How does NVIDIA seize market opportunities and maximize profits? The secret to its continued leadership lies in two key strategies.

1. 'Three Teams, Two Seasons' Strategy

On April 25, 2024, Jen-Hsun Huang personally delivered the world's first DGX H200 to OpenAI, inscribed with the message, 'Aiming to advance artificial intelligence, computing technology, and human development.' From left to right: Sam Altman, Jen-Hsun Huang, Greg Brockman.

Huang observed that no company could permanently dominate an industry. Leaders like S3, Tseng Labs, and Matrox were replaced after just one or two chip generations. This is because the industry operates at the pace of PC manufacturers, updating products twice a year—spring and autumn, with the autumn cycle being crucial.

PC manufacturers introduce new products every six months, equipped with the latest chips. They swiftly switch suppliers for faster, better components.

Chip manufacturers, including NVIDIA, take 18 months to design and launch a new chip, focusing on one at a time. However, graphics technology evolves rapidly, rendering chip designs obsolete by the time they reach the market.

Huang later announced a method to keep NVIDIA ahead: dividing the design team into three groups. The first focused on new chip architectures, while the other two developed iterative versions based on these new chips. This allowed NVIDIA to release a new chip every six months, aligning with PC manufacturers' procurement cycles.

This 'Three Teams, Two Seasons' strategy relies on early technological decisions, such as the 'Virtual Object' architecture and backward-compatible drivers. Now, NVIDIA can launch three times as many chips, each with a development cycle of no more than six months, ensuring the latest products are available during peak sales seasons.

Even if competitors release slightly better products, PC manufacturers stick with NVIDIA, knowing new chips will arrive within six months, eliminating the hassle of changing drivers.

NVIDIA's rapid iteration means 'competitors are always one step behind.' As Huang later said, 'Time is the first element of any product.'

2. 'Delivering the Whole Cow' Strategy

'Delivering the Whole Cow' refers to using nearly every part of a cow, not just premium cuts. In 'The Innovator's Dilemma,' Clayton Christensen notes that industry development follows predictable patterns.

Startups introduce disruptive innovations initially inferior to market leaders but positioned in the low-end market. Market leaders often overlook this less profitable segment, focusing on products maintaining strong profit streams.

However, disruptive innovations create new applications, and startups iterate and innovate faster. Eventually, startups have more powerful products, while incumbents face trouble too late.

Huang was determined to prevent this fate for NVIDIA. He knew competitors would struggle to surpass NVIDIA's high-quality chips due to the capital and engineering talent required. Instead, he saw a threat from low-cost participants.

Inspired by Intel's product line defending against low-end market attacks, NVIDIA repurposed rejected components as weaker, cheaper versions of its main products, profiting without costly R&D.

This created a new derivative product line defending against low-cost chip competitors. While the cheaper line might incur losses, sales of high-end products more than compensated.

II. Flat Corporate Structure

Huang envisioned an ideal organization from scratch, choosing a flat structure enabling independent employee action. This structure also weeded out low performers reliant on instructions.

Traditional Pyramid Corporate Structure

Huang believed the traditional pyramid structure—with top executives at the top, middle managers in the middle, and grassroots employees at the bottom—hindered excellence. Instead, he reshaped NVIDIA into a computer stack or short cylinder.

NVIDIA's structure contrasts sharply with most American companies, where CEOs typically have few direct subordinates. In the 2010s, Huang's senior management team included over 40 executives, later exceeding 60, each reporting directly to him.

This promoted a culture of transparency and knowledge sharing. With few layers between executives and grassroots employees, everyone could assist in problem-solving and prepare for potential issues.

NVIDIA's openness was impressive. While most CEOs have eight or nine people in large executive meetings, Huang's conference room was always full.

To reduce information ambiguity, Huang spent considerable time communicating with employees, ensuring everyone understood the company's strategy and vision. This level of transparency was typically reserved for 'C-level employees' in most companies.

'By having a large number of direct reports rather than a one-on-one approach, we've flattened the company, enabling rapid information transmission and empowering employees,' Huang said. 'This structure is well-designed.'

III. Two Key Work Methods

Work must be fast, efficient, and always serve customer interests.

1. Speed of Light Work Approach

In September 2023, Kelli Valade, CEO of Denny's, and Jen-Hsun Huang unveiled a plaque at Denny's in San Jose, California, where NVIDIA was founded in 1993.

Since NVIDIA's inception, Huang has insisted on the 'Speed of Light' work ethic.

This is measured against physical limits, not compared to other companies or NVIDIA's past achievements. Each project is broken down into basic tasks with target completion times excluding delays, waiting periods, or downtimes.

'The Speed of Light allows you to enter the market faster. Even if it can't prevent competitors from surpassing you, it makes it difficult for them to do so,' Huang said.

If subordinates set goals based on past achievements or competitors, Huang would reprimand them, preventing internal corruption where employees intentionally adjusted timelines for career advancement.

The 'Speed of Light' concept ensures NVIDIA never engages in deliberate delays.

In 1997, NVIDIA launched its third-generation product, NV3, known as RIVA 128, exemplifying the 'Speed of Light' project planning.

Traditionally, chip production begins with a physical prototype. Once complete, software engineers write drivers, and the chip design is optimized. To save time, Huang decided to develop drivers before the RIVA 128 prototype was complete, revolutionizing the process and shortening the production cycle by nearly a year.

2. The Mission is the Boss

Large companies often have multiple business units managed by competing executives, slowing progress through internal politics. When things don't go smoothly, companies lay off employees to meet budget targets, even top talent.

This fuels short-term thinking and internal information hoarding.

As Huang said, 'You want a company large enough to do business well but as small as possible,' not bogged down by excessive management and processes.

NVIDIA Headquarters in Santa Clara, California, USA

He told employees their ultimate boss was the mission itself, meaning all decisions benefit customers, not executives' careers.

Huang grouped employees by function—sales, engineering, operations, etc.—viewing them as a general talent pool rather than divided by business units or departments.

Managers were instilled with the idea that they didn't 'own' employees but should be accustomed to their movement between task groups, avoiding internal friction in large companies.

'Managers don't feel empowered by having large teams. At NVIDIA, you gain power by accomplishing amazing work,' Huang said.

This made NVIDIA faster and more efficient, with quicker decision-making. Employees, regardless of level, had a say in every decision, based on the quality and value of information and data, not leaders' promotional needs or forcing obedience.

IV. Three Key Management Methods

NVIDIA employees' daily work is guided by three key management methods.

In January 2024, Jen-Hsun Huang danced the yangko dance at the annual meeting at the Shanghai office.

1. 'Five Top Priorities' Email Method

It's impossible to discuss everything in meetings. For a large, widely distributed organization, Huang needed a way to stay informed, ensuring everyone had the right priorities.

He required employees at all levels to send emails to their superiors and executives, detailing the five major tasks they were working on and recent market observations, including customer pain points, competitor activities, technological advancements, and potential project delays.

To filter these emails, Huang asked each department to label them by topic.

'Five Top Priorities' emails became an important feedback channel for Huang, allowing him to detect early market changes that might be obvious to junior employees but not yet perceived by him or senior management.

'I'm looking for faint signals. Strong signals are easy to capture, but I want to detect and intercept them while they're still faint,' Huang said.

These emails also provided insight into new market opportunities. When Huang was interested in a new market, he shaped his strategic thinking nearly in real-time through these emails.

For instance, after perusing several emails detailing the "Five Top Priorities," which discussed the latest trends in machine learning, Jen-Hsun Huang realized that the company was not advancing swiftly enough to fully capitalize on this burgeoning market. He promptly informed his team that they required more software engineers.

Jen-Hsun Huang was known for responding to emails within minutes of receiving them and expected his employees to reciprocate within 24 hours with well-thought-out and data-driven responses.

"Nothing should be neglected. No issue should be allowed to fester. You must act promptly and drive these matters forward."

Array

Array

Array

Array

Array

Array

Array

Array

Array

Array

Array

Array

Array

Array

Array

Array

Array

Array

Array