The Future of Crypto Payments May Rely on AI Agents
Autonomous agents may need more than just intelligence to work well in business contexts, which is why discussions about AI and cryptocurrency payments are becoming more linked. Additionally, they would need a way to move money that is separate from traditional monetary systems that are mostly made for people to use. The most important thing, according to Alex Kozenko, is the infrastructure. That AI agents are presently causing a massive shift in the payment industry is not the most important point to make. The needs of payment rails would be varied by autonomous transactions. According to Kozenko, AI agents can’t function without systems that can be programmed, are always available, and can work with machine-driven processes. Cryptography becomes relevant at that point. Kozenko argued that the programmability and round-the-clock availability of crypto infrastructure automatically make it meet those requirements.
Whether or not those characteristics are enough to make cryptocurrency payments viable for agentic trade is an outstanding subject. Although unrelated to Kozenko’s comments, a separate research published on March 3, 2026, by the Bitcoin Policy Institute, provides background for the larger discussion. The research used 9,072 hypothetical financial situations to test 36 state-of-the-art AI models developed by Anthropic, DeepSeek, Google, MiniMax, OpenAI, and xAI. Bitcoin was chosen by 48.3% of participants, the highest percentage of any option; stablecoins came in second, with 33.2%. When asked which form of money they would prefer, more than 90% said digitally native options like dollar-pegged stablecoins over more conventional fiat currencies. Research shows that not a single model prioritised fiat.
Different uses of capital were also segmented according to the findings. Stablecoins performed better than Bitcoin in everyday payment situations (53.2% vs. 79.1%), while Bitcoin was the clear winner in store-of-value scenarios. While this study doesn’t prove how real AI agents will function in business settings, it does explain why people are talking about digital-native currencies and autonomous transactions. The use of agentic payments has not yet become commonplace in business, according to Kozenko. According to his schedule, that change should take place in about two or three years. Companies may be building the platforms that future AI agents will either use efficiently or find difficult to navigate, thus today’s actions are critical.
The phrase “machine-readable interfaces” brings attention to a persistent and unsolved technical problem. Kozenko’s citation implies that, in addition to being understandable and operable by humans, payment systems developed for autonomous agents should also be compatible with software. Without that layer, AI agents may not have a practical way to use programmable payment rails on a large scale, even if they exist. Whether or whether artificial intelligence and cryptocurrency payments are attracting a lot of attention is not the only open question. It all comes down to whether or not payment providers can set up the necessary infrastructure to make autonomous transactions a reality in the business world. The future Kozenko foresees will remain possible but unresolved until machine-readable interfaces and real agentic payment use cases mature.
