Jensen Huang, Co-founder & CEO at NVIDIA
4.6/5 Rating
Technology
$1M+/mo
Approx. $215.9 Billion ARR

Jensen HuangCo-founder & CEO

In this interview, NVIDIA co-founder and CEO Jensen Huang discusses the geopolitical landscape of AI, the co-existence of open-source and closed models, and the massive capital expenditure shift powering the artificial intelligence revolution. Huang addresses trade dynamics with China, critiques regulatory over-correction, and highlights how open models strengthen global cybersecurity. He outlines NVIDIA's headcount strategy and shares how personal trials and a relentless focus on execution build leadership resilience.

Jensen Huang

Jensen Huang

Co-founder & CEO

NVIDIA

NVIDIA

Founder Stats

  • Technology
  • Started 1993
  • $1M+/mo
  • 50000+ team
  • Santa Clara, California, United States

About Jensen Huang

Jensen Huang is the co-founder and CEO of NVIDIA, the world's leading company in programmable graphics and artificial intelligence computing. Born in Taiwan, Huang moved to the United States as a child and graduated from Oregon State University and Stanford University with degrees in electrical engineering. Under his leadership from day zero in 1993, NVIDIA transformed from a 3D graphics startup into the world's most valuable technology enterprise, powering the global AI infrastructure.

Interview

July 24, 2026

This is interview 2 of 2 with Jensen Huang.

Jensen Huang has been featured on CEO Insider 2 times. Explore the full collection.

Q

How would Chinese export controls on AI models affect the U.S. technology sector?

Question 1 of 18
Jensen Huang

I hope that they continue to keep it open. The Chinese market produces groundbreaking research, and half of the world's AI researchers are Chinese. We want to make sure that we can all share ideas as quickly as possible to advance the technology safely.

0
Q

Should the U.S. government restrict or ban Chinese AI models like Kimi?

Question 2 of 18
Jensen Huang

I hope not. The industry and science need open models for cybersecurity, economic security, and national security. Companies need to be able to control their intellectual property and data, which requires open frameworks.

0
Q

Is there a security risk in American companies using Chinese open-source models?

Question 3 of 18
Jensen Huang

No, that is a misconception. When you download an open model, you operate it within a harness inside a secure sandbox. This sandbox maintains privacy, security, and access control, allowing you to fine-tune and guardrail the model as you desire.

0
Q

Why is it essential to regulate AI applications rather than the raw technology?

Question 4 of 18
Jensen Huang

The raw technology is dual-use and must be allowed to advance quickly to become safer and more secure. Regulators should focus on specific applications, such as how AI is utilized in medicine, autonomous transportation, or local services.

0
Q

What is the current status of NVIDIA's sales to the Chinese market?

Question 5 of 18
Jensen Huang

Our sales to China are approximately zero today due to export controls. I have told our investors not to expect any sales. If licensing rules change and we are allowed to return, it would be a great honor to support that market.

0
Q

Why has the market repeatedly misunderstood the impact of open-source Chinese models like DeepSeek?

Question 6 of 18
Jensen Huang

Investors panicked and dumped chip stocks, thinking open models reduce the value of hardware. In reality, great open models lower the barrier to entry, which drives massive adoption. More adoption requires more compute, data centers, and NVIDIA chips.

0
Q

What is the purpose of NVIDIA's open-source model, Nemotron?

Question 7 of 18
Jensen Huang

We built Nemotron for companies and industries that cannot outsource their core intelligence due to sovereignty, regulatory, or privacy concerns. It is designed to be adapted for specific custom skills rather than competing directly with frontier labs.

0
Q

Why should companies avoid building their own custom AI models from scratch?

Question 8 of 18
Jensen Huang

Building and hosting your own AI is incredibly clumsy, expensive, and time-consuming. Most companies should rent pre-built intelligence from frontier labs like OpenAI or Anthropic because it is more convenient and cost-effective than running it themselves.

0
Q

How do open-source models strengthen global cybersecurity defense?

Question 9 of 18
Jensen Huang

Closing models creates a single point of failure. Opening models allows cybersecurity experts worldwide to inspect, test, and patch vulnerabilities, creating a decentralized and diverse defense system that makes global digital infrastructure far more secure.

0
Q

Why do you reject the narrative that AI will cause massive job losses?

Question 10 of 18
Jensen Huang

Historically, technology increases productivity, which creates more economic opportunities and jobs. For example, automating scan analysis increased the number of patients radiologists could see, which actually drove a twenty percent increase in radiologist employment.

0
Q

How does the current AI chip demand cycle differ from historical semiconductor boom-and-bust cycles?

Question 11 of 18
Jensen Huang

This cycle is not driven by seasonal consumer demand. It is an industrial computing platform shift. We are building the foundational intelligence infrastructure layer of the world, which requires the semiconductor industry to expand five to ten times.

0
Q

Why will the next generation of software companies be capital-expenditure heavy?

Question 12 of 18
Jensen Huang

Traditional software was capital-expenditure light because code was written by humans. AI software requires massive computing clusters to manufacture tokens and process reasoning. Generating intelligence requires building physical data centers and investing in hardware.

0
Q

Why do you believe that a market bubble is unlikely in the next five years?

Question 13 of 18
Jensen Huang

Supply is heavily constrained by physical limits like land, grid power, chip packaging, and construction workers. These physical bottlenecks prevent over-building, giving the market plenty of time to absorb capital expenditures productively.

0
Q

How do you manage NVIDIA's headcount to remain as small as possible?

Question 14 of 18
Jensen Huang

It is about strategy and maximizing the impact of limited resources. We have fifty thousand employees and want to grow to only seventy-five thousand in ten years. We focus our team single-mindedly on high-impact projects.

0
Q

What is the significance of the "token" in the economics of modern computing?

Question 15 of 18
Jensen Huang

A token is an embedding of knowledge and intelligence. Unlike static numbers, the mathematics behind tokens get smarter and more capable over time. As tokens become more useful, their value and profitability continue to grow.

0
Q

Why do you believe that pain and suffering are essential to achieving professional greatness?

Question 16 of 18
Jensen Huang

No athlete becomes great without setbacks, losing, and practicing when nobody is watching. Pain and suffering build character, resilience, and focus. They teach you how to manage stress and stay calm when challenges are at their highest.

0
Q

Why do you choose not to wear a watch or manage your life with an outlook calendar?

Question 17 of 18
Jensen Huang

Now is the most important time. I refuse to let a watch or a calendar manage my life. By focusing one hundred percent of my attention on the present moment, I can dedicate myself completely to the people in front of me.

0
Q

Why does the United States remain the land of dreams for global entrepreneurs?

Question 18 of 18
Jensen Huang

It is the greatest country in the world because our open discourse, freedom, and vibrant system allow anyone with desire to make a great living. The environment naturally attracts the brightest minds to collaborate and build the future.

0

Video Interviews with Jensen Huang

Jensen Huang says the AI doomers have it wrong

Jensen Huang says the AI doomers have it wrong

Jensen Huang says the AI doomers have it wrong

Nvidia GTC Taipei 2026 Keynote

Nvidia GTC Taipei 2026 Keynote

Cite This Interview

Use this interview in your research, article, or academic work

Related Interviews

Travis Kalanick, Founder & CEO at Atoms
Technology

Travis Kalanick

Founder & CEO at Atoms

Not Publicly Disclosed
2000+

In this interview, Atoms founder and CEO Travis Kalanick shares his vision for automating the physical world through industrial robotics and AI. Kalanick reflects on…

Read Interview
David Baszucki, Founder & CEO at Roblox
Technology

David Baszucki

Founder & CEO at Roblox

Approx. 400 Million USD
3100+

In this interview, Roblox co-founder and CEO David Baszucki discusses the platform's evolution into a major digital human co-experience platform. Baszucki outlines the value of…

Read Interview
Garry Tan, President & CEO at Y Combinator
Technology

Garry Tan

President & CEO at Y Combinator

Not Publicly Disclosed
100+

In this interview, Y Combinator President and CEO Garry Tan outlines the new paradigms of software development driven by vibe coding and AI agents. Tan…

Read Interview
Ted Sarandos, Co-CEO at Netflix
Technology

Ted Sarandos

Co-CEO at Netflix

Approx. 3.75 Billion USD
16000

In this interview, Netflix Co-CEO Ted Sarandos discusses the platform's ten-year journey in India and its global expansion strategy. Sarandos explains how Netflix matches content…

Read Interview
Brian Schimpf, Co-founder & CEO at Anduril Industries
Technology

Brian Schimpf

Co-founder & CEO at Anduril Industries

Approx. $183.3 Million
7000+

In this interview, Anduril Industries co-founder and CEO Brian Schimpf discusses the shift in modern warfare toward AI-powered, software-first defense technologies. Schimpf shares the origin…

Read Interview
Amanda McMaster, Interim CEO at Boston Dynamics
Technology

Amanda McMaster

Interim CEO at Boston Dynamics

Not Publicly Disclosed
1500+

In this interview, Boston Dynamics interim CEO Amanda McMaster explains the commercial transition of the pioneer robotics firm from research laboratory to industrial deployer. McMaster…

Read Interview