NVIDIA CEO Jensen Huang on AI by 2025 An Era for Grounding and Growth
In an extensive conversation looking back at 2025, NVIDIA's COO Huang painted a picture of the impressive development within the AI industry, mainstream narratives of the future, and often-missed pragmatic ground-level realities. He spoke profusely of leaps in grounding and reasoning that have dented initial concerns about AI-generated gibberish-making AI a very reliable assistant for human experts in medicine and law, just to mention a few.
On AI Impact on Jobs Task versus Purpose
Huang tackles the mass job displacement narrative by juxtaposing a framework of task versus purpose. While he claims that AI handles some tasks-coding or studying scans-it does not take away the purpose of jobs, that being solving problems or diagnosing a disease.
The Radiologist Example: Radiologists were supposed to become obsolete by predictions of AI, but their ranks have grown. AI has enabled radiologists to read more scans and diagnose disease better, thus creating a stronger demand for their skills.
New Industries and Labor Shortages: AI is creating new industries and jobs, especially with new chip plants, supercomputer plants, or AI factories. Huang explains how robotics and automation should act as remedies to monstrous labor shortages in manufacturing and trucking.
The Robot Repair Industry: Huang predicts that with a billion robots in existence, the largest repair and maintenance industry will be created in history.
Geopolitics, China, and the Vital Role of Open Source
Huang's framework visualizing the AI technology stack as a five-layer cake-energy, chips, infrastructure, AI models, and applications-is used in his exposition on the criticality of open source and its nuanced relationship with China.
Open Source is Essential: He offers dire caution that without open source in place, startup innovation, as well as programs in higher education and established industries like healthcare and manufacturing, would be suffocated. It is paramount to have policies that do not occur to destroy this innovation flywheel.
A Nuanced View on China: The complete decoupling of China is nothing but a naive proposal in Huang's perspective. Huang proposes a balanced relationship, where China stands as an adversary but as a partner too. He mentions that China forms a significant market for American technology, as well as the largest provider of manpower for open source projects, notably DeepSeek, which has directly fed American AI labs.
Against the Doom and AI Bubble Question
Huang is especially concerned with "doomer" and dystopian sci-fi narratives, which in his opinion are unhelpful and can result in tax-policy-making that hampers innovation. He finds an equal lack of nuance in "AI bubble" discussions.
More Than Chatbots: According to him, it is too simplistic to link the health of the entire AI market to the revenues of some chatbot companies. NVIDIA's business in the areas of accelerated computing for autonomous vehicles, digital biology, financial services, and robotics is already a multi-billion dollar business sector, independent of revenues from generative AI chatbots.
A Fundamental Shift in Computing: Huang states that the major cause for growth is industry-wide transition from general-purpose computing to accelerated computing, a trend that arose due to the end of Moore's law. AI is an application of this wider shift, not the primary one.
Insane Demand: He cites the desire for computing capacity as a universal truth across all industries, a testament to the real, fundamental value.
"Name one startup that said, 'Nah, we don't need it.' They are all dying for computing capacity."
What's Next The Science and Robotics "ChatGPT Moment."
Huang envisions that moves into fundamental multimodality, long context, and synthetic data will give rise to several industries experiencing their own "ChatGPT moment."
Digital Biology: Generative AI for digital biology holds great promise in Huang's particular enthusiasm: protein generation, chemical understanding, and building foundation models for cells and proteins.
Robotics and Self-Driving Cars: He expects reasoning advances to unlock giant leaps in robotics and self-driving vehicles regarding dealing with out-of-distribution scenarios they aren't explicitly trained for. Unlike self-driving, he is optimistic about humanoid robotics moving way stronger, essentially building on the framework and forward technology of contemporary AI.
