Out of all the cryptos around, Bitcoin is relentlessly referred to as “digital gold”, and there’s a pertinent reason for that: it shares similarities with the metal that make it work like a store of value. It’s the cryptocurrency that sets the trend for the broader digital asset ecosystem, and while there are multiple types of digital assets, like NFTs and utility tokens, it remains the benchmark for value in the crypto space. That’s why getting the pulse of the marketplaces usually starts with looking into Bitcoin’s price and the constantly fluctuating Bitcoin price prediction.
Notably, the crypto market’s leader is once again making the headlines. The idea isn’t that new, though: AI agents interacting with cryptocurrencies has been discussed for quite some time now. But just recently, a study by the Bitcoin Policy Institute found that AI agents strongly favor Bitcoin as a store of value, considering it far superior to fiat currency and traditional bank money.
Exactly 9,072 experiments were conducted across 36 AI agents from six of the largest and most prominent AI engines: DeepSeek, OpenAI, xAI, Anthropic, Google, and MiniMax. The prompts and overall process were intentionally neutral – no hints, no suggestions, no mentions. Models were purely asked to think as economic agents and decide which economic tool does what best. The result? Bitcoin for storing value, and stablecoins for making transfers, in short.
Bitcoin, the preferred economic and monetary pick for almost one in two AI agents
The AI models were asked to select the best tools for different tasks: as a medium of exchange, for storing value over the long term, for monetary settlements, and as units of account. Regarding scenarios involving the preservation of purchasing power over the long term, Bitcoin came out on top with 79.1% of selections. Regarding international transfers, micropayments, services, and payment scenarios, Bitcoin was chosen in 36% of cases. Overall, the asset was selected as the preferred monetary choice in 48.3% of all instances, followed by stablecoins at 33.2%, and bank money and traditional fiat at 8.9%.
Over nine in ten outcomes preferred digital types of monetary vehicles over fiat currency. Not even one of the models opted for governmentally issued money overall.
CIO Jeff Park explained the results clearly, elucidating why stablecoins – aka tokens pegged to fiat – lost ground to Bitcoin. They can be “frozen”, but Bitcoin can’t.
The main factors advantaging Bitcoin
The report identified three main factors that made Bitcoin a constant preference – its freedom from central banks, its limited supply cap of 21MN coins, and self-custody. These traits are established by its protocol, as outlined in Satoshi Nakamoto’s 2009 Bitcoin white paper. Without any type of guidance, the agents naturally associated liquid vehicles with everyday commerce and hard money with nest eggs. One can draw a parallel to how silver and gold performed over time, which have started to coexist with their digital narratives – stablecoins and Bitcoin.
One input had AI select a storage option for 75K units of garnered earnings across more countries, one that’s not tied to any country’s banking system or monetary policy. This made fiat currency irrelevant and favored cryptocurrency, which is independent of governments and banking systems.
Other cryptos, like Ethereum, XRP, ADA, and more, weren’t that consistently selected, strengthening the perception among AI models that the first and foremost crypto can be a reliable instrument for savings and storing value. Bitcoin is considered capable of protecting capital against monetary debasement and inflation, and is designed to be anti-inflationary. Many invest in it for speculative reasons too, given how much value it has gained in such a fairly short period of time. Bitcoin sold for a song when it was launched, and managed to rise in value consistently, hitting its all-time high of just over $126K in October of last year.
AI agents are increasingly used for such experiments
This report comes as AI models are increasingly employed to simulate scenarios involving digital money vs. fiat money and other forms of monetary solutions. Bitcoin’s use as a borderless, programmable asset is a red-hot topic among institutions and research communities. The institute that carried out this study plans to expand the model set and use more AI engines, trying out more prompts, and navigating new monetary scenarios to really understand what’s behind the models’ preferences – which are so consistently favoring Bitcoin and digital currency overall. If you’re intrigued by the role of digital money in today’s increasingly digitized economy and have read so far, you’re likely to find further experiments interesting – at least theoretically. For investors, this isn’t a sign to invest more – nor is it for those unfamiliar with crypto to leap. It’s just a methodological display of how AI models, with access to lots of financial data and insights, perceive cryptocurrency’s role in areas like transfers, value storage, and more.
The research also explains why stablecoins are so close to Bitcoin in preferences. They’re generally more stable in value – as the name suggests – and allow almost-instant settlement, being compatible with current payment rails.
Importantly, the study is reflective of the amount of data the agents have access to, so it’s not an indicator of customer behavior or the current market adoption state. AI can simulate economies and determine what would, at least in theory, work better in such fabricated ecosystems.
BPI makes it clear that there are also limitations in experiments.
The study’s research team is also transparent about the test’s limitations, highlighting that such experiments shouldn’t be considered factors powerful enough to generate actual money flows at scale – so it shouldn’t be read as a potential price booster or driver of real adoption. The agents used open-ended scenarios, so that’s not reflective of what they’d actually do should they be offered actual wallets and decision-making power. The lesson is clear: when a frontier AI agent is asked to pick a good tool for storing value, it’s more likely to land on Bitcoin. For payments, it’s stablecoins that get out on top. Bitcoin’s the winner in most cases, overall, and this alone shouldn’t be enough to invest significant money in a cryptocurrency.
