Nvidia CEO Jensen Huang Rejects AI Doom Scenarios, Calling Existential Threats 'Complete Nonsense'

In an interview with Axios's Mike Allen, Nvidia co-founder and CEO Jensen Huang pushed back against widespread warnings of existential AI hazards and extreme labor disruption, labeling claims that the technology will destroy humanity or eliminate half of all jobs as "complete nonsense." Huang argued that current evidence demonstrates AI boosting productivity across sectors such as radiology, legal services, and manufacturing, which ultimately increases demand for human workers. He cautioned policymakers against falling for "science fiction" narratives that could lead to over-regulation, emphasizing that the primary economic threat comes from failing to adopt AI technology rather than adopting it too quickly.

Huang also addressed recent market volatility fueled by the emergence of powerful Chinese open-source models like DeepSeek and Kimi. He argued that Wall Street fundamentally misunderstands the dynamic, explaining that accessible open models drive broader AI experimentation and market diffusion. Increased AI usage across society creates higher demand for proprietary frontier models like OpenAI and Anthropic, while simultaneously driving sales for Nvidia's compute hardware and data center infrastructure. Huang emphasized that the technology ecosystem requires both closed cloud services for enterprise convenience and open models for sovereignty, custom enterprise IP, and robust cybersecurity.

Looking at the broader semiconductor and computing landscape, Huang stated that the industry is only at the beginning of a multi-year buildout, dismissing concerns of an impending tech bubble or cyclical bust in the next five to ten years. He noted that computing is undergoing a fundamental platform shift toward generated intelligence, necessitating an artificial intelligence layer built on top of existing global infrastructure. To meet future demand, Huang estimated that global semiconductor manufacturing capacity will need to grow five to ten times larger over the next decade, with current growth primarily constrained by physical limitations in power, memory, land, and supply chain capacity.