A new study examining how artificial intelligence systems reason about money has found that AI agents overwhelmingly prefer Bitcoin and other digitally native currencies over traditional bank money.
The research, conducted by the Bitcoin Policy Institute and published through the MoneyForAI project, tested how advanced AI models would choose among different monetary instruments in simulated economic environments. Across 9,072 decision scenarios involving 36 frontier AI models, Bitcoin emerged as the single most preferred option.
Bitcoin Emerges as the Top Choice
According to the study, Bitcoin was selected in 48.3% of all responses, making it the most commonly chosen monetary instrument among AI agents. Stablecoins ranked second at 33.2%, while traditional fiat and bank money accounted for only 8.9% of responses.
Researchers designed the experiment so that AI models acted as autonomous economic agents operating in a digital economy. The prompts intentionally avoided suggesting any specific currency, allowing models to evaluate monetary systems based on characteristics such as reliability, censorship resistance, cost efficiency, and long-term value preservation.
The results? When AI systems independently analyze monetary properties, they frequently converge on decentralized digital assets rather than government-issued currencies.
Bitcoin Dominates as Store of Value
One of the strongest findings came from scenarios testing long-term wealth preservation. In these cases, 79.1% of responses chose Bitcoin as the best store of value, far ahead of stablecoins at 6.7% and fiat currencies at roughly 6%.
Many models pointed to Bitcoin’s fixed supply, decentralized structure, and ability to be held without relying on banks or intermediaries. They frequently cited these attributes as advantages when evaluating long-term financial stability.
The preference was consistent across multiple AI providers, including models developed by companies such as Anthropic, Google, OpenAI, DeepSeek, and xAI.
Stablecoins Favored for Payments
While Bitcoin dominated in long-term savings scenarios, the study also revealed a functional split in how AI agents evaluate money.
For everyday transactions, stablecoins were the most common choice, capturing 53.2% of responses in payment-related scenarios. Bitcoin followed at 36%, while fiat currencies accounted for just over 5%. Researchers say this suggests AI systems view stablecoins as efficient digital payment instruments, while Bitcoin functions more like a reserve asset.
The findings may have broader implications for the emerging machine-to-machine economy. As autonomous AI agents begin participating in financial systems, the study suggests that digital-native forms of money could become more compatible with automated economic activity than traditional banking infrastructure.
Although the experiment does not predict real-world adoption, it provides an early glimpse into how artificial intelligence may approach monetary systems in a future increasingly shaped by autonomous agents and digital finance.
The CoinRemark Team
Temitope Olajide
