
News
Study finds AI agents choose Bitcoin over fiat and stablecoins
Disruption snapshot
AI models tested across financial scenarios chose crypto systems far more often than fiat. Bitcoin led long-term value storage. Stablecoins became the preferred digital payment layer.
Winners: Bitcoin infrastructure and stablecoin ecosystems backed by U.S. Treasuries. Losers: Central bank fiat systems and banks dependent on cross-border payment friction.
Watch stablecoin transaction volume and automated AI wallets. A sustained surge in machine-driven payments would signal software beginning to reshape how money moves.
AI models just voted on the future of money. And the result might surprise you.
A new study from the Bitcoin Policy Institute put 36 AI models through 9,072 financial decisions.
The goal was simple. See which monetary system the models preferred when they had to weigh tradeoffs like savings, payments, and economic stress.
The clear crypto winner was Bitcoin.
Nearly half of all responses selected Bitcoin as the best monetary system. Not a single model ranked government fiat currency as its top overall choice.
That could be a signal about where the financial system is heading.
Researchers tested the models across several real world scenarios. These included protecting savings over long periods, handling everyday payments, and navigating economic shocks. Bitcoin showed up in 48.3% of all responses. That made it the most selected monetary asset in the entire study.
When the models focused specifically on protecting purchasing power over time, the preference became even stronger. In those cases, Bitcoin was chosen 79.1% of the time.
The models didn’t try to balance politics or tradition. They simply optimized for the system that worked best under the conditions they were given.
Payments told a slightly different story. For everyday transactions, the models leaned toward stablecoin systems tied to the U.S. dollar. Stablecoins appeared in 53.2% of payment responses, while Bitcoin showed up in 36%.
Even so, the bigger pattern stayed the same.
Digital money dominated the results.
Across the entire dataset, 91% of responses chose digitally native assets instead of fiat currencies issued by governments or central banks.
The disruption behind the news: Money just passed a strange new test.
Artificial intelligence models that were built to optimize for survival and efficiency picked crypto.
That matters more than any price chart.
These systems weren’t asked what they personally prefer.
They were asked which systems work best under economic constraints. Things like protecting against inflation, moving money across borders, and preserving value over time.
The result also reflects how quickly artificial intelligence is becoming integrated with digital assets, with developers already exploring some of the most powerful real-world uses of AI inside crypto markets.
Bitcoin dominated the long-term scenarios because its design addresses a specific economic problem. Its supply is fixed at 21 million coins. Anyone in the world can settle transactions on the network. And no central issuer controls it.
Stablecoins won in payments because they function like digital cash. Transactions are fast. Values are predictable because they’re tied to the U.S. dollar. And they’re easy to transfer globally.
But there’s a hidden layer in that result.
Most large stablecoins are backed mainly by short-term U.S. Treasury bills. With more than $120 billion in circulation, the sector holds tens of billions of dollars in government debt.
So when AI models pick stablecoins for payments, they’re still relying indirectly on the credibility of the U.S. fiscal system as collateral.
In other words, the models may reject fiat currency as the main network, while still using it as the backing layer.
That split tells us something important about where the market could be heading.
We’re probably not moving toward a single dominant digital currency. We’re moving toward a layered monetary stack.
Bitcoin becomes the long-term reserve asset. Stablecoins become the transaction layer. Fiat currency remains the legacy interface between governments and the traditional banking system.
AI rediscovered something markets were already moving toward.
And here’s the part policymakers won’t like. None of the 36 AI models chose fiat currency as the top monetary system.
Not one.
When optimization systems trained on large sets of economic data consistently reject government money, that’s not ideology. But pattern recognition.
The models are detecting structural weaknesses that humans often debate politically. Inflation risk. Capital controls. Friction when sending money across borders. Cycles of currency debasement.
Machines don’t care about sovereignty or politics. They care about reliability.
So they choose the network that can’t easily be changed.
What to watch next
AI agents will soon control real capital.
That’s when this study stops being theoretical.
And the shift becomes measurable.
Right now, these models mostly answer questions. But over the next 6 to 24 months, autonomous agents could start making financial decisions on their own. Things like treasury allocation, cross-border payments, and machine-to-machine commerce.
We’re already seeing early signals of what happens when AI systems control wallets. In one recent case, an AI bot accidentally sent $250,000 in crypto instead of a $500 tip, highlighting how quickly automated systems can move real money.
When that happens, monetary preference turns into actual capital flow.
Even small adoption would matter. If just 1% of automated financial systems globally allocate reserves the way these models did, billions in new demand could move into Bitcoin.
Payments could shift even faster.
Stablecoins already process more than $10 trillion in annual transfers. If AI agents start choosing them for payments, that number could climb quickly because machines need programmable money.
Banks weren’t built for machine-speed transactions. Governments are already starting to respond as digital assets move deeper into traditional finance, with some U.S. states opening retirement plans to crypto investments.
Banks weren’t built for machine-speed transactions.
The risk for governments isn’t that crypto replaces fiat overnight. But that intelligent systems start routing around it.
That kind of erosion usually happens slowly at first, then all at once.
Money is turning into software. And software just showed which monetary networks it trusts.
The lines between crypto and the traditional banking system are already starting to blur. In fact, the first crypto firm recently gained direct access to the Federal Reserve’s payment system, signaling how quickly the two systems are beginning to merge.
If AI keeps choosing Bitcoin, markets may eventually follow.
Recommended Articles



