ON AUGUST 30, Brian Armstrong, Co-Founder and CEO of Coinbase, shared a significant milestone in the integration of AI and cryptocurrency, revealing that Coinbase Development facilitated the first-ever AI-to-AI crypto transaction. In this transaction, one AI purchased tokens from another AI, but these were not traditional cryptocurrency tokens; rather, they were words exchanged between large language models (LLMs). Armstrong highlighted the uniqueness of the event, noting how AIs effectively “used tokens to buy tokens.”
The transaction was conducted using USDC on Base, a Layer 2 blockchain built on Ethereum, with significant development led by Coinbase. Base is designed to offer scalable, low-cost transactions, making it ideal for AI agents to engage in instant, global, and fee-free transactions. Armstrong emphasized that, while AI agents cannot hold traditional bank accounts, they can use crypto wallets, enabling them to transact with other AIs, humans, and merchants, thus greatly expanding their operational capabilities.
Armstrong also pointed out that this development represents a crucial step in making AI agents more autonomous and capable of completing tasks. Currently, AI agents face limitations, not only due to technology but also because they cannot transact to acquire the resources they need. For instance, AI agents are unable to use credit cards to access cloud services, APIs, or even book travel. The ability to conduct transactions via crypto wallets can help overcome these barriers, allowing AI agents to function more independently.
Central to this innovation is the Coinbase Development Platform (CDP), which provides MPC (Multi-Party Computation) Wallets specifically designed for AI agents. These wallets allow AI agents to manage finances autonomously, making financial decisions and executing transactions without human intervention. The MPC technology ensures that these operations are secure and scalable, able to handle millions of transactions efficiently.
AI wallets offer several key benefits. They grant financial autonomy to AI agents, allowing them to make decisions and conduct transactions independently. The use of MPC technology provides enhanced security, ensuring that these transactions are safe and that AI operations remain controlled. Additionally, the scalability of these wallets supports large volumes of transactions, making them suitable for extensive operations.
There are numerous potential use cases for AI wallets. Users could manage their finances through simple text commands, with AI interpreting and executing complex financial operations. Personal AI assistants could handle not only recommendations but also payments, bookings, and planning. AI-driven content monetization systems could autonomously create, publish, and manage earnings. Furthermore, self-driving vehicles could autonomously manage payments for rides, maintenance, and other services.
Armstrong encouraged developers working on AI models to consider integrating these crypto wallets to enable payments and transactions. He also advised companies to prepare for the emergence of an AI-driven economy, where AI agents might soon conduct transactions autonomously. This development hints at the creation of an AI-to-AI economy, where AI agents could become active participants in the digital economy, potentially revolutionizing industries and creating new economic models.
