In a recent in-depth interview, NVIDIA CEO Jensen Huang engaged in a candid conversation spanning over an hour, covering topics such as the future of AI, competition between China and the U.S., open-source versus closed-source models, job anxiety, and technology regulation. As the leader of the world's most valuable company, his views are both straightforward and insightful.
The following is a summary of the conversation:
1. On Competition with China: Openness Matters More Than Blockades
When asked about reports that China might tighten export controls on AI models and semiconductors, Huang's first reaction was, "I hope not."
He pointed out an overlooked fact: "Half of the world's AI researchers are Chinese, many of whom still live in China. They have made some truly breakthrough research achievements in AI."
He believes openness is crucial for technological progress: "We want to ensure that everyone can share ideas as widely and quickly as possible to advance this incredible endeavor. Use technology in ways that are beneficial and safe for society."
When asked whether U.S. companies should be allowed to use Chinese AI models, his answer was succinct: "Of course."
He explained the reasoning: "There's a misconception that certain shortcuts are connected to China. But once a model is downloaded, you can fine-tune it, improve it, set up guardrails, and run it in a sandbox. These sandboxes are secure and protect privacy and security."
Regarding sales to China, Huang stated clearly: "Currently, our sales to China are nearly zero. I've told all investors not to have any expectations for sales in China. If the U.S. issues a license and the Chinese market wants us back, it would be our honor. But until then, I can only assume sales are zero."
2. On Open-Source Models: Kimi, DeepSeek, and Competition
When asked if the rise of Chinese models like Kimi threatens U.S. AI companies, Huang offered a counterintuitive view:
"These Chinese models are excellent. The market misunderstands the impact of DeepSeek and Kimi. I believe the great success of open models is a good thing for the entire industry."
Why? Because demand is key.
"When AI becomes more powerful, there are more uses. When demand increases, more NVIDIA computers need to be sold, more data centers need to be built, more services need to be provided, and technology will spread to more industries. Models can bring about widespread applications, leading to tremendous growth."
He dismissed the misconception of "open models versus closed models": "The people most likely to upgrade to using Anthropic and OpenAI are those who are already using AI. They want to have it more conveniently. Open models are incredibly cheap, easy to get started with, and convenient for trying AI. This is the best thing we can do."
Huang painted a clear picture: "The world will have both closed and open models simultaneously. We should use closed models as much as possible, and we should also use open models when necessary to build our own personalized AI."
He offered an important judgment: "For many countries and companies, they must build their own intelligent systems. They cannot outsource core intelligence to a third party. You should rent intelligence resources from cloud services as much as possible, but you cannot outsource your own wisdom. You should build your own customized AI."
This directly echoes Palantir CEO's warning about "don't give up your intellectual property": "No one should outsource their core team. No one should outsource their intelligence. No country should do that."
3. AI Creates Jobs, Not Destroys Them
Huang's stance on this issue is extremely firm:
"Saying that AI will destroy half of America's jobs is complete nonsense. All facts, all evidence points to the exact opposite direction."
He backed it up with data: "The number of radiologists worldwide has increased by about 20%. The reason is that scanning work has been automated by AI, but they can treat more patients. Because there are too many backlogs, more radiologists are needed. The number of paralegals has increased by about 10%. Manufacturing jobs have increased by about 50% because these computers must be connected to AI data centers."
His core point is: "Those who use AI will take the 'jobs' of those who don't. Therefore, we must ensure that AI is introduced into all industries as quickly as possible. Anything that restricts this is untimely."
4. On 'Doomsday Claims': Complete Nonsense
Huang was unsparing in his criticism of AI doomsday scenarios:
"The end of humanity is coming—complete nonsense. The idea that AI will destroy half of America's jobs—complete nonsense."
"All evidence points to the exact opposite. Intuition and common sense will tell you this is completely wrong. AI will improve productivity. When productivity improves, it creates opportunities."
He believes fabricating facts is "absolutely inappropriate": "If leaders want to warn the world, they should provide solutions, not make up sci-fi stories. Saying we live in a simulation, or that the singularity is coming—these are all fabricated."
He even joked: "The closest things to real AI are R2-D2 and C-3PO. Who wouldn't want R2-D2 and C-3PO?"
Huang made a direct appeal to policymakers:
"Before listening to all the empty rhetoric, before hearing all the stories, before the imaginary visions of AI's future—none of this exists. It's all science fiction."
He believes governments need to distinguish between the technology itself and its applications: "The technology itself is dual-use. Should regulate the application of AI in specific fields like medicine and transportation. But the technology itself should develop as quickly as possible."
Regarding the idea that the Trump administration might "take a stake in leading AI companies," Huang said it's unnecessary: "The U.S. already owns a stake in NVIDIA—its success is tied to ours. Last year, we paid $10 billion in taxes. The more successful a U.S. company is, the more tax revenue and jobs it creates, and all Americans benefit."
5. On the American Spirit and Competition
Huang's belief in America runs throughout the conversation:
"America is the greatest country in the world. Not because we have fewer challenges, but because in the face of enormous challenges and differences, we manage to solve them through open discussion and a vibrant system."
"This country has freedom of speech, freedom of innovation, freedom of entrepreneurship, and the environment naturally attracts exceptional talent. I recommend that all smart people from around the world come here. This country was built by immigrants, and it will need excellent immigrants in the future."
On relations with China, he is both realistic and confident: "It's wrong to think we can hinder China's development. But China cannot stop America from moving forward either. We have outstanding talent from all over the world. The U.S. tech industry is too active, too innovative, and moves too fast. Whatever competition comes from around the world, bring it on. America is ready."
6. On the Future: AI's 'Infrastructure Moment'
Huang believes the AI industry is at a historic moment:
"In the past six months, AI has finally become truly useful. Useful AI is also profitable AI. When we have profitable AI, we want to develop more AI."
He describes a massive shift: "In the past, computers were all pre-recorded content. In the future, we will generate intelligence on top of that. No need to search, just ask a query, and AI will give you the answer."
"We need more computers. The semiconductor industry likely needs to be 5 to 10 times larger than it is now over the next decade. Right now, it's too small, which is why everything—memory, storage, optical interconnects, packaging, chips—is so scarce."
He also touched on the future of AI agents: "In the future, we will have 100 billion agents running all the time. So the number of computers we need will grow significantly."
7. Personal Reflections: Pain and Hardship Are Necessary
Towards the end of the interview, Huang shared some personal insights.
On success: "The path to greatness lies in experiencing a lot of pain and hardship. Great athletes are great because of lots of practice, especially when no one is watching. They face many setbacks and lose many times. That's what truly improves your craft, builds character, and enhances resilience."
On time management: "I refuse to let Outlook control my life. I refuse to let a watch control my life. If I'm late, someone will tell me. In the meantime, I'm 100% here. Now is the most important moment."
On weekends: "I wake up early, play with the dogs, work until lunchtime, have lunch, work until dinner, the kids come over, we cook dinner together, have cocktails, play with the dogs—a perfect day."
Huang's conversation reveals a different side of him than what media often portrays—not a techno-utopian or a doomsayer, but an engineer who speaks with data and logic. In his view, AI is neither a threat nor magic, but an infrastructure that needs to be widely applied. To win this race, the U.S. relies not on blockades, but on openness, enthusiasm, and action.
Comments