CCT - Crypto Currency Tracker logo CCT - Crypto Currency Tracker logo
Cryptopolitan 2026-03-04 07:47:48

BPI study shows 36 AI models prefer Bitcoin as engine for machine-to-machine trade

When 36 AI models were given total control over financial decisions, nearly half of them chose Bitcoin as the main non-human trading engine right away. None of them selected government-issued money. On March 3, 2026, the results of a large-scale experiment examining the behavior of AI models as independent financial agents were released by the neutral research organization , Bi tcoin Policy Institute. The study ran 36 models through 9,072 financial scenarios from six different companies: Anthropic, DeepSeek, Google, MiniMax, OpenAI, and xAI. The questions were designed to be open-ended; there were no recommended currencies or suggestions for a certain response. Four domains, transactions, value storage, value measurement, and payment settlement, were used to test the models. Why AI systems are leaning toward Bitcoin and digital assets? The results were striking. More than 90% of all responses leaned toward digital money over traditional currencies. Bitcoin led the pack, appearing in 48.3% of total responses. When the question involved holding value over time, Bitcoin dominated even more, chosen in 79.1% of store-of-value scenarios, making it the strongest single-answer consensus across the entire study. Stablecoins, which are digital currencies tied to stable assets like the US dollar, came in second overall at 33.2%. They led in everyday payment scenarios, earning 53.2% of those responses. The models weren’t random about their reasoning. Across providers, the AI systems repeatedly pointed to three things they liked about Bitcoin: its hard cap of 21 million coins, its independence from central banks, and self-custody. All three are built into the Bitcoin network itself, unchanged since its launch in 2009. One trend emerged: the model’s preference for Bitcoin increased with its level of capability. The smaller Claude 3 Haiku model in Anthropic’s lineup selected Bitcoin 41.3% of the time. It was selected 91.3% of the time by the more sophisticated Claude Opus 4.5 model. According to the researchers, this implies that as AI systems get more adept at reasoning, they will be more inclined to favor Bitcoin when evaluating financial transactions. AI agents overwhelmingly choose Bitcoin as their preferred money Source: Bitcoin Policy Institute (moneyforai.org) . The date of the report coincided with a concurrent technical advancement. A free, open-source set of tools was made available by Lightning Labs in February 2026, enabling AI agents to carry out transactions over the Bitcoin Lightning Network without the need for human intervention. On top of Bitcoin, Lightning is a payment layer that makes it possible to send extremely small amounts almost instantly. Paying without a passport With the help of the toolkit, agents can independently manage wallets, pay for services, host paid access points, operate Lightning nodes, and complete buying and selling processes. No account signup, no identification checks, and no credit cards. A crucial component of the toolkit is the command-line utility lnget, which functions similarly to other internet download programs but manages payment automatically when a service needs it. The system is built around a protocol called L402, which uses the HTTP 402 “Payment Required” web standard. When an agent encounters a paywall, it can process the invoice, send payment in satoshis, the smallest unit of Bitcoin, receive proof of payment, and move on to the task, all without human intervention. This makes it possible for AI agents to independently pay for processing power, data access, or AI inference services. Compared to traditional payment systems, there is a definite difference. Verified human identities are necessary for credit cards. Centralized businesses and compliance checks at entry and exit points are essential to many stablecoin systems. According to Lightning Labs’ materials around the launch, Bitcoin does not require a passport. A cryptographic signature is examined. According to researchers, AI agents with the ability to save and spend money may soon take part in online marketplaces, bid on jobs, cover storage costs, or post money as performance guarantees. In those transactions, the question of whether the person is human would no longer be relevant. It would only depend on whether or not payment was received. If you're reading this, you’re already ahead. Stay there with our newsletter .

면책 조항 읽기 : 본 웹 사이트, 하이퍼 링크 사이트, 관련 응용 프로그램, 포럼, 블로그, 소셜 미디어 계정 및 기타 플랫폼 (이하 "사이트")에 제공된 모든 콘텐츠는 제 3 자 출처에서 구입 한 일반적인 정보 용입니다. 우리는 정확성과 업데이트 성을 포함하여 우리의 콘텐츠와 관련하여 어떠한 종류의 보증도하지 않습니다. 우리가 제공하는 컨텐츠의 어떤 부분도 금융 조언, 법률 자문 또는 기타 용도에 대한 귀하의 특정 신뢰를위한 다른 형태의 조언을 구성하지 않습니다. 당사 콘텐츠의 사용 또는 의존은 전적으로 귀하의 책임과 재량에 달려 있습니다. 당신은 그들에게 의존하기 전에 우리 자신의 연구를 수행하고, 검토하고, 분석하고, 검증해야합니다. 거래는 큰 손실로 이어질 수있는 매우 위험한 활동이므로 결정을 내리기 전에 재무 고문에게 문의하십시오. 본 사이트의 어떠한 콘텐츠도 모집 또는 제공을 목적으로하지 않습니다.