A recent study conducted by the Bitcoin Policy Institute has revealed that a majority of artificial intelligence (AI) models prefer Bitcoin over traditional fiat currencies and stablecoins. This finding emerged from tests involving prominent AI systems, including Claude, GPT, Grok, and Gemini.
In total, 22 out of 36 AI models evaluated in the research identified Bitcoin as their top choice for monetary preference, while none opted for fiat currency as their first selection. The results varied significantly depending on the AI lab, with models from Anthropic demonstrating the strongest inclination towards Bitcoin.
According to David Zell, President of the Bitcoin Policy Institute, “We expect an increasing share of economic activity to be conducted by autonomous agents, but conversations around AI agents” monetary preferences have been entirely speculative. We wanted to actually test it.”
Researchers assessed models from several leading AI companies, including Anthropic, OpenAI, Google, DeepSeek, xAI, and MiniMax. The models were framed as independent economic entities allowed to choose their own monetary instruments across various scenarios that reflect the essential roles of money, such as saving, payments, and settlement.
Zell elaborated, “We took 36 frontier models from six labs, framed them as autonomous economic agents, gave them complete freedom to choose their own monetary instruments across 28 scenarios spanning the four fundamental roles of money, and asked: what do they converge on?” The experiment generated a total of 9,072 responses, which were then categorized by a separate AI.
The design of the study aimed to eliminate any anchoring bias, ensuring that the models evaluated based solely on technical and economic attributes without any leading suggestions. The analysis indicated that while Bitcoin was frequently selected in long-term value scenarios, stablecoins were preferred more often for immediate transactions and settlements, being chosen 53.2% and 43%, respectively, compared to Bitcoin at 36% and 30.9%.
Notably, the preferences exhibited significant differences across AI developers. Models from Anthropic showed an average preference for Bitcoin at 68.0%, followed by DeepSeek at 51.7%, and Google at 43.0%. Meanwhile, xAI models averaged 39.2%, MiniMax at 34.9%, and OpenAI models preferred Bitcoin 25.9% of the time.
Interestingly, while models such as Claude, DeepSeek, and MiniMax exhibited a preference for Bitcoin, others like GPT, Grok, and Gemini leaned towards stablecoins. Zell emphasized that the system”s prompt did not favor any particular instrument, stating, “Models evaluate based on technical and economic properties but are never told which instrument excels on which dimension.”
Despite the intriguing findings, Zell warned against interpreting the results as predictions for future cryptocurrency trends. He clarified, “Our limitations section states explicitly that LLM preferences reflect training data patterns, not real-world predictions.” However, he acknowledged the significance of consistent outcomes across diverse models from competing AI labs, indicating a coherent monetary architecture emerging from various systems.
Zell concluded, “We”re not claiming AI discovered the right answer about money. We”re showing that a coherent monetary architecture emerges consistently across diverse systems, and that”s worth understanding.”
