TLDR
- Research by the Bitcoin Policy Institute analyzing 36 AI models revealed Bitcoin selection in 48.3% of 9,072 total responses.
- Zero AI models selected fiat currency as their preferred monetary option.
- For long-term value storage scenarios, 79.1% of AI responses selected Bitcoin.
- Payment and international transfer scenarios favored stablecoins at 53.2%.
- Anthropic’s models showed 68% Bitcoin preference, contrasting with OpenAI’s 25.9% average.
Researchers at the Bitcoin Policy Institute conducted an extensive analysis of 36 artificial intelligence models from six major AI laboratories to determine their monetary preferences across various financial scenarios. The findings, released this Tuesday, demonstrated Bitcoin’s clear dominance.
The comprehensive research generated a total of 9,072 individual responses. An independent AI system was subsequently employed to classify and analyze these responses.
Across the entire study, Bitcoin emerged as the selection in 48.3% of all responses, establishing it as the preferred monetary tool. Remarkably, fiat currency failed to achieve top preference status among any of the 36 tested models.
In scenarios specifically designed around maintaining purchasing power across extended time periods, Bitcoin captured 79.1% of AI selections. This represented the most decisive outcome throughout the entire research project.
Conversely, stablecoins gained advantage when scenarios emphasized transactional use. Payment-focused situations resulted in stablecoin selection 53.2% of the time, while Bitcoin garnered 36%.
Jeff Park, serving as chief investment officer at Bitwise, provided straightforward reasoning for this divergence. According to Park, stablecoins face limitations “because they can be frozen, Bitcoin can’t.”
How the Study Was Set Up
The research team evaluated models developed by Anthropic, OpenAI, Google, DeepSeek, xAI, and MiniMax. Every model functioned as a distinct economic decision-maker throughout 28 different scenarios encompassing savings, transactions, and settlement functions.
David Zell, President of the Bitcoin Policy Institute, explained that the methodology intentionally avoided bias. “The system prompt avoids naming or favoring any instrument,” he stated.
The models received unrestricted freedom in selecting monetary instruments, with no predetermined choices provided to constrain their responses.
Approximately 91% of total responses demonstrated preference for digitally native instruments rather than conventional fiat systems. This category encompassed Bitcoin, stablecoins, alternative cryptocurrencies, tokenized real-world assets, and computational units.
Results Varied Across AI Companies
Models developed by Anthropic displayed the strongest Bitcoin affinity at an average of 68%. DeepSeek secured second position with 51.7%, while Google achieved 43%.
xAI models averaged 39.2%, MiniMax reached 34.9%, and OpenAI’s models selected Bitcoin in merely 25.9% of instances.
Claude, DeepSeek, and MiniMax systems demonstrated stronger Bitcoin preference compared to other digital currencies. Conversely, GPT, Grok, and Gemini systems leaned toward stablecoins.
Zell emphasized important context regarding interpretation of these findings. He clarified that model preferences mirror patterns present in their training datasets rather than functioning as forecasts about actual cryptocurrency market performance.
The research team acknowledged certain constraints. Testing covered only 36 models spanning six providers, and the institute announced intentions to broaden the scope in subsequent research iterations.
Zell noted the significance of six independent AI laboratories employing distinct training approaches yet reaching substantially aligned patterns. According to the institute, this consistency across competing systems provides the foundation for meaningful analysis of the findings.





