NVIDIA GTC 2026: Jensen Huang Flags AI Inference Inflection Point as Data Centers Head Toward Trillion-Dollar Scale

NVIDIA has dominated AI computing for more than 20 years through its CUDA and GPU ecosystem. The generative AI boom initially propelled demand for training large models, but enterprise adoption is now shifting toward model inference and AI agents. CEO Jensen Huang described this at GTC 2026 as an “inference inflection point” whose scale will shape investment across the data-center and semiconductor supply chains.

During his GTC 2026 keynote, Huang said global demand for AI computing power was surging and that the AI computing market could approach $1 trillion by 2027, while data centers were also moving toward trillion-dollar scale. NVIDIA plans to use industry-specific libraries, its CUDA software stack and next-generation computing platforms to address long-term computing demand as the market shifts from training to large-scale inference.