A new analysis of 380 trillion AI tokens reveals that stock markets are paying an "AI Premium" - companies with high exposure to AI consumption outperform those with low exposure by about 0.64% per week. The study, led by Yale economist Aleh Tsyvinski, shows that AI's financial impact extends far beyond the technology sector, already reshaping valuations across industries.
The researchers examined one of the largest datasets of real-world AI usage ever studied, covering January 2024 through April 2026. The data came from OpenRouter, a platform that routes requests to more than 400 AI models, representing 2% of monthly global AI consumption. The granularity allowed them to track how growth in AI usage correlates with stock returns, labor trends, and the emergence of agentic AI.
The AI Premium in financial markets
The researchers constructed a weekly "AI Factor" to measure global AI consumption growth. Stocks that moved in tandem with AI usage earned higher future returns. The difference between the most and least exposed companies - what they term the AI Premium - averaged 0.64% per week. That premium is concentrated in the United States, Europe, and other developed markets where investors and companies are tied closely to frontier AI development.
The effect is weaker in China and emerging markets, where connections to leading-edge AI infrastructure are less direct. "The equity markets now reward the proximity of companies to the most frontier models in the United States and Europe," said Yukun Liu, associate professor of finance at the University of Rochester and a co-author of the study.
Sectors benefiting beyond tech
The AI Premium is not limited to software firms. Retail, consumer durables, and capital-intensive industries such as manufacturing also show positive exposure. The study defines exposure as market expectations: investors believe these companies will gain from AI-driven productivity, even if they are not yet seeing direct benefits.
Nicola Borri, associate professor of finance at Luiss University and a co-author, said that the premium is driven by intensive AI use - professional, paying customers using advanced models with sophisticated prompts - rather than casual experimentation with free tools. "Despite the widespread and fast adoption of AI tools by everyday users, the AI premium is mostly determined by the exposure to the frontier AI consumption by sophisticated and professional users," Borri said.
Impact on workers and the rise of agentic AI
The study also mapped AI consumption to occupational data. Jobs involving non-routine tasks - persuasion, teaching, communication - showed more positive exposure to AI. In contrast, roles heavy in routine analytical work, including scientific analysis and lab tasks, had negative exposure. "Our analysis suggests that interactive job skills will be rewarded while analytical skills get penalized with the rise of AI," said Tsyvinski.
The dataset also captured the fast growth of agentic AI - systems that can operate autonomously to complete tasks. In 2024, agentic models accounted for a small share of tokens. By 2026, they made up more than half. Early evidence suggests that companies' exposure to the agentic economy is beginning to influence equity valuations, the researchers said.
Why this matters for finance professionals
The findings signal that AI exposure is becoming a measurable factor in asset pricing. For portfolio managers and analysts, tracking a firm's AI consumption patterns - not just its sector label - could offer an edge. AI for Finance training can help professionals interpret these shifts and apply them to investment decisions. Meanwhile, CFOs evaluating their company's AI strategy might consider how the market rewards proximity to frontier models. An AI Learning Path for CFOs provides a structured way to build that understanding.
Your membership also unlocks:
