Jensen Huang Says NVIDIA’s Core Strength Is Turning Electrons Into Tokens as AI Agents Drive Software Growth

Generative AI primarily produces text and images in response to prompts, while agentic AI can plan and execute tasks. Physical AI extends those models into robots and factories. NVIDIA has expanded from a GPU supplier into an end-to-end AI infrastructure platform spanning chips, systems, software and data centers. Its strategy will shape the global computing industry and Taiwan’s supply chain.

At a Taiwan employee meeting held at the T17 and T18 sites in Taipei’s Beitou-Shilin Technology Park on May 27, 2026, NVIDIA Chief Executive Jensen Huang said AI was moving toward agentic systems and physical robotics. He also said he expected the company’s market capitalization to continue rising. NVIDIA plans to build its Taiwan headquarters at the site and hire another 3,000 people. Reports did not disclose the investment amount or the date when construction would begin.

NVIDIA is the global leader in AI accelerators and data-center computing platforms, and spending on its chips has become central to capital expenditure by cloud giants including Amazon, Microsoft and Google. CEO Jensen Huang says agentic AI can perform tasks autonomously and generate vast numbers of tokens, directly converting computing capacity into cloud-service revenue. The company's outlook is therefore a key gauge of whether AI investment remains robust.

NVIDIA on May 20, 2026, reported results for its fiscal 2027 first quarter ended April 26. Revenue rose 85% year over year to $81.615 billion, exceeding FactSet's estimate of $78.91 billion. Adjusted earnings per share were $1.76, also topping the $1.75 forecast. Data-center revenue climbed 92% to $75.2 billion, while the company projected second-quarter revenue of $91 billion, plus or minus 2%.