Jen-Hsun Huang: We all agree that "those who do not use AI will be replaced," and appropriately opening up the Chinese market can make the United States stronger.

NVIDIA founder Jensen Huang said at the "Winning the AI Race" forum held at The Hill and Valley Forum on Capitol Hill and the Silicon Valley Forum how AI is reshaping jobs, infrastructure and U.S.-China technology wrestling, and gave great recognition to Trump's energy and reindustrialization policies. (Synopsis: Trump once considered "splitting Huida" and found that Nvidia was too strong, Huang Jenxun responded with a rainbow fart) (Background supplement: Huang Jenxun: I use AI to make myself smarter like this, and I don't worry about AI robbing human jobs) If you don't use AI, you'll be replaced by people who use AI, and I think we're all pretty sure about it. With the lights on the podium, NVIDIA CEO Jensen Huang used this unconcealed warning to set the tone for the "Winning the AI Race" panel in the summer of 2025. From jobs to infrastructure, from Trump policies to chip export bans, Huang dismantles complex technologies and geopolitics into readable roadmaps that are both computing power corporate strategies and the current state of global capital flows. Huang Jenxun reminds: AI has no mercy on employment Faced with the old question of whether AI will eliminate jobs, Huang Jenxun pulled the answer to 100% job transformation. He called AI "the greatest technical equalizer of all time", even without a programming background, AI can teach people to write programs and design. Inside NVIDIA, every engineer is assisted by AI, and the workload is even greater, because ideas and experimentation emerge when monotony is eliminated. He roughly estimates that AI will automate about half of routine tasks, but at the same time expand human influence and create more new jobs. This "augmentation, not replacement" argument sounds optimistic, but it is lethal. Jensen Huang concluded: "If you don't use AI, you will be replaced by people who use AI." What he meant was to urge companies and individuals to make up for AI penetration as soon as possible, otherwise all career chains would be reshuffled. Industrial Imagination of AI Factories For AI imagination to land, a new type of "AI factory" consisting of electricity, data centers, and high-speed networks is required. Huang estimates that this will be a market of "trillions of dollars", and the key is the NVIDIA GB200 Grace Blackwell super chip: the inference performance is 30 times higher than the previous generation, and it can be accelerated upward through software iteration; Through the CUDA ecosystem and vertically integrated solutions, NVIDIA is no longer just selling chips, but providing an end-to-end platform to strengthen customer retention. Huang Jenxun also proposed the concept of "physical AI": in the future, all industrial companies will operate traditional factories and AI factories at the same time, such as Tesla while building cars, optimizing production lines with AI models and even autonomous vehicle driving. When all moving objects are connected to AI, the old division of labor will be rewritten, and whoever can build the data and computing power infrastructure first will have the dominance of the next decade. Praise Trump, talk about US-China AI confrontation Geopolitics is another line of dialogue at the forum, Huang Jenxun rarely publicly praised Trump for "grasping the importance of AI and energy on the first day of office", strongly agreed with Trump's policy tone, and advocated re-industrialization and high-tech local manufacturing. Huang estimates that the domestic AI industry in the United States is expected to add trillions of dollars in output in the next four years, and emphasizes that "without energy, there is no new industry", throwing out the signal of energy thirst. The U.S. chip export controls on China have put the industry at a crossroads, and Huang believes that comprehensive restrictions on the export of chips will be "counterproductive" and will only prompt China to accelerate self-research. He revealed that he successfully lobbied the Trump administration to agree that NVIDIA will sell H20 AI chips to China and maintain China's dependence on the US technology stack, which is more in the long-term interest than cutting off. The "clearance" deal is seen as a realignment between Washington's economy and national security. Huang also named the progress of Chinese AI labs in the field of open source models, such as DeepSeek, Kimmy, etc., believing that open source is conducive to global innovation and consolidates the US technology stack as a standard. In other words, he hopes to maintain the interdependence of Silicon Valley and the Chinese market through "semi-openness", rather than decoupling, but to make American AI go faster. AI talent needs capital The most scarce AI race may not be chips, but talent. There is a story in the market that "AI researchers signed a contract of more than 100 million US dollars", and Huang Jenxun did not deny it. He admits that he personally audits the salaries of the company's 42,000 employees and uses machine learning to ensure fairness. He would rather raise his salary than lose key engineers, because "take care of the employees and the rest will fall into place." With increasingly powerful tools and infrastructure, he believes that a small team of 150 people can produce OpenAI-level results that impact traditional large enterprises. The model of small teams and huge computing power makes capital chase top experts more than institutions, and the value of talents rises as a result. From "100% jobs reshaped" to "mega-AI factories," Huang uses a series of logics to circle the four pillars of the AI era: skills, infrastructure, geopolitics, and human capital. The energy policy and re-industrialization vision of the Trump era have unexpectedly become the fuel for AI hegemony in his mouth. The limited opening to China shows that Washington is taking into account the softening of Silicon Valley's industry. For everyone in Huang's audience, the question is no longer whether AI will come, but how to keep up with the acceleration of AI tools. Related reports Tether invests in empire with 120 companies: dabbled in bitcoin mining, AI, Taiwan XREX, Juventus FC OpenAI cooperates with Oracle to build 4.5 GW data center, Stargate AI total power breaks 5 GW AIT: The United States and Taiwan negotiate to build a spaceport in Pingtung Jiupeng, and the suborbital flight to Houston is only 2.5 hours 〈Huang Jenxun: We all agree that "people who don't use AI will be replaced", Properly opening up the Chinese market can make the United States stronger" This article was first published in BlockTempo's "Dynamic Trend - The Most Influential Blockchain News Media".