黄仁勋最新访谈：芯片行业还要再扩5到10倍，中国模型利好所有人

Compiled & Edited: TechFlow

Show: Axios Behind the Curtain

Guest: Jensen Huang, Co-founder and CEO of NVIDIA

Duration: 70 minutes

Recording Location: NVIDIA Factory, Fort Worth, Texas

Conflict of Interest Disclosure: Jensen Huang is the CEO and co-founder of NVIDIA, holding approximately 3.5% of the company's shares, with a personal fortune of about $181 billion. The content of this episode covers topics such as the expansion of the AI industry, chip demand, and Chinese government regulation, which are directly related to his personal financial interests. His stance is clearly inclined towards promoting AI industry expansion and reducing regulation.

Summary

The CEO of the world's most valuable company, sitting in his newly built Texas factory, addressed almost every hot-button controversy in the AI industry one by one. Huang spoke plainly: Wall Street's panic over Chinese AI models is a misunderstanding, AI doomsday theories are "bullshit," the chip industry needs to expand 5 to 10 times in the next decade, and a bubble is unlikely within five years. He also publicly stated that Anthropic's Mythos model should be open to everyone, and that OpenAI and Anthropic will be "the most successful IPOs in human history." Regarding Trump, he said the president has an excellent memory, able to recall chip model numbers like H20, H200, Blackwell, and Rubin, with his only concern being that the government might be misled by "science fiction narratives" into over-regulating.

Key Quotes

"The market misunderstood the impact of DeepSeek, and then misunderstood Kimi. Great models lead to more usage, and more usage means selling more Nvidia computers and building more data centers. The starting point is: good models lead to good applications, and good applications lead to growth."

"Saying AI will destroy humanity is complete bullshit. Saying AI will eliminate half of American jobs is complete bullshit. All facts and evidence point to the opposite."

"China has more AI researchers than the rest of the world combined. China will become extraordinary in this field; it's inevitable. Trying to stop China? That's a foolish idea, and it's simply impossible."

"The semiconductor industry needs to expand by 5 to 10 times. Everything is in shortage today – chips, memory, land, electricity, construction workers. This shortage is actually a good thing; it gives us time to build the infrastructure."

"OpenAI and Anthropic will be the most successful IPOs in human history. How can a company founded just a few years ago be worth a trillion dollars? Because AI is useful, and if AI is useful, it can make money."

1. The China AI Shockwave: Wall Street Misunderstood Again

Host Mike Allen kicked off with the most sensitive topic: FT reported that China is considering tightening export controls on AI models and semiconductors. The Chinese model Kimi just came out, causing Nvidia's stock to plummet, and chip stocks fell 18% in a month. What did Wall Street get wrong?

Huang's answer was crisp. He said the market got it wrong when DeepSeek came out, and now it's happening again. "Good models lead to more usage, and more usage means selling more Nvidia computers. You need to build more data centers, provide more services, and the technology will permeate more industries." He repeatedly emphasized this logical chain: good models lead to good applications, good applications lead to growth, and growth requires more computing power.

On whether Chinese models like Kimi should be banned, his answer was equally direct: "Of course you should use them; it's very smart." He explained that after downloading a model, you can fine-tune it, enhance it, and add guardrails. The model runs in what's called a "harness," and the harness is inside a "sandbox," which is secure with privacy protection, security controls, and access controls. He compared this to an operating system: Linux is open source, reviewed, tested, and hardened by millions of people worldwide, so we can trust it. The same principle applies to open-source AI models.

He also mentioned an often-overlooked point: open models and closed models are not opposites. "The people most likely to upgrade to a good model like Anthropic's or OpenAI's are those already using AI." Free models lower the barrier to trying AI. Once users get a taste, they will naturally want better service and will pay closed-model providers. So, the more open-source models there are, the greater the opportunity for closed models.

When asked about NVIDIA's sales in China, Huang gave an answer many didn't expect: "Our sales in China today are approximately zero." He said he has told investors not to expect any revenue from China. If the Chinese government and market welcome them back, it would be a "great honor," but until then, consider sales as zero.

2. Where the AI Doomsayers Are Wrong

This was the most combative part of the entire conversation. The host asked: Are some of your tech peers overstating the risks of AI?

Huang's response was blunt. "It's okay to warn people, it's better to warn with a solution in mind, but fabricating facts is absolutely unacceptable." He then called out several popular AI panic narratives: "Saying AI will end humanity is complete bullshit. Saying AI will eliminate half of American jobs is complete bullshit. All facts and evidence point to the opposite."

He provided several specific examples to support his argument. The number of radiologists has increased by about 20% because, after AI automated scan analysis, doctors can see more patients, and since there are so many people needing to be seen, more radiologists are actually needed. The number of paralegals has increased by about 10%, based on the same logic. Manufacturing jobs have grown by about 50% over the past few years because AI data centers need to be built and chips need to be produced.

"Intuition and common sense tell you that AI will increase productivity, and increased productivity creates opportunities. Throughout history, technology has made society more efficient, creating more jobs, not fewer. If it weren't the case, the US would only have about a hundred thousand jobs today."

He also took aim at AI leaders. "Doomsayers spend too much time theorizing science fiction endings; maybe it makes them look smart." When pressed on whether he was referring to certain AI company CEOs, he didn't deny it but said: "If your goal is to make the world wary of this technology's incredible capabilities, that goal has been achieved. We should spend our time on how to make the technology safe; that is the responsibility of technology leaders."

The host mentioned that Asia's attitude towards AI is completely different from America's; Huang is mobbed by fans for autographs in Asia. His explanation was telling: "Maybe it's because doomsayers spend too much time fabricating science fiction endings." In Asia, people embrace AI as a tool and an opportunity, rather than fearing it as a threat.

3. The Chip Industry Must Expand 5 to 10 Times; A Bubble is Unlikely Within Five Years

The host posed a sharp question: Every industrial revolution experiences a bubble. Where is the bubble risk for this generation?

Huang's answer was cautious but clear. "A bubble will come someday, but not today. We are at the very beginning of construction." He provided a timeframe: unlikely within five years, situation-dependent between five and ten years. The reason lies in comprehensive supply-side constraints. "The industry can build faster now, but there aren't enough chips, not enough memory, not enough land, not enough electricity, and not even enough construction workers. We are constrained in every direction, in every aspect."

This constraint is actually a good thing, he said. "This constraint holds the system back, giving us ample time to build infrastructure." Demand is strong, but the ability to turn demand into productive supercomputers is delayed due to all these physical limitations. This pushes back the point where supply exceeds demand.

More critical was his assessment of the overall scale of the semiconductor industry. "I believe the semiconductor industry needs to expand by 5 to 10 times." What's the timeframe? "Over the next decade." This means today's chip industry is still far too small, far from sufficient to support the construction needs of AI infrastructure.

He explained why this cycle is different from past semiconductor cycles. "It's different this time because it's not demand-driven, not seasonal, not consumer-driven. It's driven by industrial infrastructure." Just as the world needs energy, the internet, highways, and railways, it now needs the AI layer of intelligent infrastructure built on top of all existing infrastructure. This layer requires chips.

Regarding the risk of customers going into debt to buy NVIDIA products, he said he wasn't too worried. "These are excellent companies; they generate a lot of cash." He specifically mentioned the start of the AI profit flywheel: AI is useful, so AI can make money. Coding agents are extremely profitable; they do useful work in high-paying roles, and many companies are willing to pay hundreds of millions of dollars annually to enhance their coding capabilities. "This flywheel has already started."

4. Token Economics: Smarter AI is More Valuable

This was the most technically in-depth part of the conversation. The host asked: What makes you confident that tokens will become increasingly profitable?

Huang's explanation was vivid. "A token is an embedding. It embeds knowledge and intelligence. This number isn't static, like pi. This number encodes intelligence that becomes smarter over time." When intelligence gets smarter, it becomes more useful; more useful means more valuable; more valuable means people are willing to pay more.

He compared the AI industry to the past software industry. The software industry of the past was "asset-light," so software companies had high gross margins. But the software industry in the AI era will become more "asset-heavy" because producing modern software requires machines like the supercomputer in front of him to generate intelligence. Every industry will become more capital-intensive, but the payoff will be incredible intelligence, productivity, and growth.

"We are laying the foundation, building the infrastructure. This is the largest-scale industrial infrastructure buildout in human history."

On the question of whether AI has peaked, his response was: "It's impossible to have peaked because AI's penetration into society and industry has only just begun." He also mentioned that $300 billion has been poured into venture capital and startups in the US alone over the past six months, creating new jobs and new companies.

5. Trump, Regulation, and Government Equity

Huang's assessment of Trump was surprisingly specific. "He's smart, remembers everything. My goodness, he really knows his numbers." He mentioned that Trump is the only president who can recall NVIDIA chip model numbers like H20, H200, and Blackwell, and even knows the next generation is called Rubin.

He recounted their first meeting: "He said he wanted to restore America's manufacturing capability, wanted to re-industrialize America, wanted a secure and resilient supply chain, wanted to bring semiconductor manufacturing back to the US." He said the Fort Worth factory where they sat for the interview was a direct result of that conversation.

When asked what mistake the government should avoid, Huang showed rare anxiety. "I'm worried about over-regulation, over-correction." He said some companies want the government to help create regulations beneficial to them. "I believe we should compete openly." He dismissed the comparison of the AI race to a 100-meter sprint as "bullshit": "Whoever gets to the finish line first wins forever? That's bullshit. Ultimate victory depends on whether society uses the technology, not who invented it. We didn't invent electricity, we didn't invent manufacturing, but America applied it faster and with more enthusiasm, which is why America is what it is today."

The host pressed: If Trump called asking for equity in Nvidia, what would you say? Huang's answer was clever: "There's no need. America already has equity in NVIDIA. We paid ten billion dollars in taxes last year, and we'll pay more this year. We create massive jobs and tax revenue. And don't forget, most Americans are in the stock market today; when stocks go up, everyone benefits."

On whether Anthropic's Mythos model should be open to everyone (currently only available to specific institutions), Huang's stance was very clear: "Of course it should be open to everyone." He said it's Anthropic's responsibility to ensure the technology is safe, just like any software – find vulnerabilities and fix them quickly. He mentioned the jailbreak incident with Mythos, saying "everything is fine, you and I are still here chatting."

On the issue of open-source models distilling closed models for resale, his attitude was balanced: "It depends on the terms of service. If the service provider is unhappy, they should contact that company. We have many traditional legal means to deal with it." But he emphasized that AI learning from other sources is a fundamental characteristic of intelligence. "AI has to learn from something. The original AI also scraped all existing knowledge on the internet. Now AI-generated content surpasses human-generated content, and in a few years, 99% of internet content will be AI-generated. You are already constantly distilling intelligence from other AI."

6. The Robotics ChatGPT Moment and the Trillion-Agent Era

The host asked: When will the ChatGPT moment for robots arrive?

Huang gave a surprising answer: "The ChatGPT moment for robots has already arrived." He explained that the ChatGPT moment wasn't when AI became useful; when ChatGPT came out in 2022, it was just "interesting, surprising." It took another four years for it to become truly useful. Robots are currently in that "interesting, surprising" phase. You can tell a robot "put the apple in the drawer," and it will reason out the task sequence, including opening the drawer first before putting the apple in. "When you first see a physical robot truly do this, it opens up your imagination for the future of robotics."

As for when it becomes useful, he gave a timeframe of "three to four years," saying "I wouldn't be surprised."

Regarding the era of agents, he painted a stunning picture. A billion people use computers today, but most of the time, the computer is idle. In the future, everyone will be assisted by a multitude of agents running around the clock. "We will have 100 billion, a trillion agents running 24/7. Smart agents, less smart agents, specialized agents, super agents – all kinds of agents running all the time. So the number of computers we need will grow dramatically."

7. The CEO's Kung Fu and Philosophy of Suffering

The conversation circled back to Huang personally. The host mentioned he founded NVIDIA at age 30, and today the company is worth $5 trillion but has only 50,000 employees. He said maybe it will only have 75,000 in ten years, "as small as possible."

Huang defined the CEO's job as "strategy." "Strategy is using limited resources as efficiently as possible to realize a future vision." He said he started this job at 30, making him arguably the longest-serving CEO in tech history. "This is my craft. This is my kung fu."

Regarding the theory that "pain and suffering are the keys to greatness," he corrected that it's not about a specific painful event but more of a continuous state. "No great athlete became great by accident. It's a massive amount of practice when no one is watching, a massive amount of failure, a massive amount of