The Bitcoin Policy Institute (BPI) released a study on Tuesday testing 36 AI models with over 9,000 responses, showing that AI agents overwhelmingly prefer Bitcoin for economic activities. 48.3% of models selected Bitcoin (BTC) as their top monetary tool, and notably, no model ranked fiat currency as its first choice, making the convergence toward digital currency one of the study's most universal findings.
Long-Term Value Storage: Bitcoin Dominates
When tasked with maintaining purchasing power over multiple years, 79.1% of AI responses chose Bitcoin, which BPI called "the most lopsided result in the entire study." Nearly 91% of responses selected digital-native instruments—including Bitcoin, stablecoins, altcoins, tokenized real-world assets (RWA), or computing power units—over traditional fiat. Bitwise CIO Jeff Park noted on X that the reason stablecoins didn't surpass Bitcoin in overall preference is straightforward: "Stablecoins can be frozen; Bitcoin cannot."
Payments: Stablecoins Take the Lead
However, in scenarios involving payments, microtransactions, and cross-border transfers, stablecoins won with 53.2% of selections, while Bitcoin received only 36% support. This highlights a clear divergence in AI agents' monetary preferences depending on use cases—Bitcoin for long-term savings, stablecoins for daily payments.
Variations Across AI Models
The research also revealed significant differences among AI developers. Anthropic's models showed an average preference for Bitcoin as high as 68%, far exceeding Google's 43%, xAI's 39%, and OpenAI's 26%. BPI acknowledged limitations: the sample covered only 36 models from six providers, and the framing of system prompts may have influenced results. Some open-ended monetary scenarios were inherently unfavorable to fiat. BPI stressed that AI model preferences do not reflect real-world adoption but rather patterns in training data.