According to CoinMeta, three economists utilized approximately 380 trillion token data points on OpenRouter to construct an index and backtest it on the U.S. stock market. The study covered the period from January 2024 to April 2026, and the results showed that the strategy of buying high-exposure stocks and shorting low-exposure stocks generated an average weekly return difference of 0.641 percentage points. Even after controlling for relevant factors, there was still a difference of about 0.56 percentage points. This method is limited by the sample period and the scope of data coverage, so its long-term effectiveness remains to be verified.