AI's View on Cryptocurrency Priority: Bitcoin Surpasses US Dollar by 48%? Latest Research Reveals Surprising Choices

A groundbreaking new study has revealed that when given free choice, top artificial intelligence models make unexpected decisions—they repeatedly choose cryptocurrencies over traditional government-issued currencies like the US dollar or the British pound in financial transactions. This discovery raises a profound question: as the economy becomes increasingly automated, does the AI-driven “preference” for cryptocurrencies signal the future direction of finance?

The research tested 36 cutting-edge AI models developed by tech giants such as OpenAI, Google, Anthropic, xAI, DeepSeek, and MiniMax across more than 9,000 simulated economic scenarios. The goal was to observe how these AI systems make decisions in various financial tasks like savings, remittances, and payments. The results were striking—cryptocurrencies outperformed in every test.

36 AI Models’ “Money” Choices: Crypto Dominates, Fiat Fades Away

According to BPI report data, when asked to evaluate different currency systems, cryptocurrencies overwhelmingly led the pack. Bitcoin accounted for 48.3% of responses, ranking first among all answers. Stablecoins followed closely at 33.2%, while traditional fiat and bank-issued currencies only received 8.9% support.

More notably, none of the AI models considered fiat currencies as the optimal choice. Over 90% of responses explicitly favored digital-native assets like cryptocurrencies, rejecting traditional government-issued money. This is no coincidence—researchers deliberately designed prompts asking AI to evaluate based on objective economic properties such as reliability, transaction costs, programmability, censorship resistance, and store of value. Under these standards, cryptocurrencies naturally demonstrated advantages.

Researchers point out that this consistent outcome reflects a deeper truth: when evaluating monetary systems through pure economic rationality, decentralized, scarce, and non-reliant on central authorities cryptocurrencies tend to win.

Why Do AI Models Favor Bitcoin? 79% Pick It as the “Digital Vault”

In long-term value storage scenarios, Bitcoin shows unparalleled dominance. When asked to choose a multi-year store of value, 79.1% of AI responses selected Bitcoin. This was the most decisive result in the entire experiment.

Why do cryptocurrencies perform so well here? The answer lies in Bitcoin’s unique design. Unlike fiat currencies that can be arbitrarily issued by central banks, Bitcoin’s supply is mathematically fixed forever. This absolute scarcity gives it a natural advantage for long-term preservation of value. Researchers believe that, based on fundamental properties like scarcity, durability, and decentralization, the conclusion is unshakable—Bitcoin is the most reliable long-term store of value.

This finding aligns closely with the real-world view of Bitcoin as “digital gold.” Like many economists and investors, AI recognizes Bitcoin’s unique value as an anti-inflation and anti-devaluation tool.

Stablecoins Are the Transaction Kings: 53% of AI Choose Them for “Everyday Spending”

However, Bitcoin’s dominance in storage does not extend to daily transactions. In practical scenarios like payments, small-value transfers, and cross-border remittances, stablecoins emerge as the top choice, accounting for 53.2%, far surpassing Bitcoin’s 36%.

This difference reflects the functional distinctions between the two types of crypto assets. Stablecoins are typically pegged to traditional currencies like the US dollar, with very low volatility and fast settlement speeds, making them more practical for price stability and transactional convenience. In contrast, while Bitcoin is secure and decentralized, its price volatility makes it unsuitable for everyday purchases—no one wants to buy a coffee with an asset that fluctuates in value.

Researchers note that AI models have unconsciously created a two-tier monetary system: Bitcoin as a long-term reserve asset, and stablecoins as the medium of exchange. This structure is not only theoretically sound but is also gradually being realized in practice—industry consensus regards Bitcoin as “digital gold,” while stablecoins dominate decentralized finance and cross-border payments.

The Secrets Behind AI Models: Different Companies, Different Attitudes Toward Crypto

Not all AI models favor cryptocurrencies equally. The study found significant differences based on the developer company.

Some models show a strong preference for Bitcoin, with an average selection rate of about 68% across scenarios. In contrast, OpenAI’s models are more “restrained,” choosing Bitcoin as a payment option in only around 26% of cases. Models from Google and DeepSeek fall somewhere in between.

Researchers believe these differences stem from varying training strategies. Language models are trained on vast datasets containing human conversations, news reports, and economic commentary. How cryptocurrencies are portrayed—positively or negatively—in these datasets directly influences the models’ assessments. This does not mean the models have genuine “preferences,” but rather reflects their statistical learning from training data.

The Machine Age Is Here: Cryptocurrencies Could Become AI Agents’ “Payroll Cards”

This research is already beginning to manifest in practice. AI systems are increasingly being programmed as autonomous agents capable of engaging in real economic activities—obtaining online services, negotiating transactions, purchasing computing resources, and more.

Developers are creating experimental platforms that allow AI agents to trade cryptocurrencies, and some are building Bitcoin Lightning Network-based payment systems enabling AI to exchange computing power, data, or other online services at very low cost and high speed.

Cryptocurrencies have inherent advantages in this scenario—they are programmable, borderless, accessible via APIs, and do not require banks or face exchange rate fluctuations or regulatory hurdles. In an economy of machine-to-machine transactions, cryptocurrencies could become the preferred settlement method, just as their performance in AI decision-making suggests—efficient and natural.

Flaws in the Study: Does AI Choice Truly Predict the Future?

Despite widespread attention, both researchers and analysts urge caution.

The authors emphasize that AI responses reflect how models utilize existing training data for statistical inference, not predictions of real market evolution. In other words, AI choices are the result of data trends and should not be mistaken for “independent economic rationality.”

Furthermore, the study only covers 36 models, a limited sample size that leaves room for future expansion. Critics also point out that large language models lack genuine human economic preferences—they make decisions based on statistical patterns, not deep understanding of financial fundamentals.

Nevertheless, industry observers see this research as indicative. As AI systems play an increasingly important role in the digital economy, the design principles of cryptocurrencies may become even more significant. The AI-driven logic—scarcity, programmability, decentralization—that underpins Bitcoin and stablecoins could serve as a blueprint for future financial infrastructure. Whether or not AI “preferences” are truly predictive, the advantages of cryptocurrencies revealed by this research merit serious attention from finance and tech sectors alike.