The United States maintained its position as the world's leading source of AI investment and AI model development in 2025, according to the 2026 Stanford AI Index, an annual assessment of artificial intelligence progress published by Stanford University's Human-Centered Artificial Intelligence institute.
US-based companies produced the majority of the world's top-performing AI models in 2025. Enterprise AI adoption in the US continued to grow, with organizations across healthcare, financial services, legal, and manufacturing deploying AI into operational workflows at scale. Among US companies that have adopted AI, productivity gains have been documented in coding, research synthesis, customer service, and data analysis functions.
The report tracks AI's progression from narrow task automation to systems capable of agentic behavior: AI that can take sequences of actions, use tools, and complete multi-step tasks with limited human oversight. This shift is reshaping how enterprises plan their AI implementation roadmaps, with multi-agent deployments becoming a key evaluation criterion for enterprise buyers in 2026.
Annual investment in AI training and deployment infrastructure in the US reached record levels in 2025, driven by hyperscaler capital expenditures and enterprise software spending. The number of publicly available AI models exceeded 1,800 globally, with US-headquartered organizations accounting for the largest share of releases.
Source: Stanford University Human-Centered AI -- https://aiindex.stanford.edu