Coinbase CEO Brian Armstrong believes blockchain and stablecoins are suitable for AI agents, considering their flexibility and adaptability to the technology. He cited underlying stablecoin characteristics, such as speed, cost of transaction, and global spread, as additional benefits that would support AI evolution.
Armstrong made the statement during a podcast interview with Nikhik Kamath, India’s largest retail stockbroker. According to him, AI agents are becoming prominent despite the absence of KYC verification, which creates significant friction in transaction rails. However, he believes KYC implementation for AI agents is jurisdictional, depending on the regulatory goals a region wants to achieve.
Notably, AI agents are spreading across various sectors of the technological ecosystem. While humans will likely continue to hold the vast majority of stored wealth and capital assets, the sheer operational velocity of machine-to-machine commerce will completely flip network usage metrics. This raises the question about how far this innovation can go in becoming the largest consumer of blockchain networks, creating demand for stablecoins, wallets, and on-chain identity before humans do.
Compared to humans, who typically execute a few financial transactions a day, a single autonomous AI agent can initiate thousands of programmatic micro-transactions per minute to buy API credits, purchase compute power, or parse paywalled data. This is happening via an ongoing transition to a machine-driven blockchain economy unfolding across core infrastructure, real-world use cases, and immediate friction points.
For instance, platforms like Squads and Crossmint are deploying agent-exclusive stablecoin accounts and virtual cards, exploring blockchain tokens’ ability to offer programmable borderless money that settles in milliseconds for fractions of a cent. Also, tech giants such as Amazon Web Services (AWS) have integrated crypto wallets into agent execution loops for fully autonomous payments, while major web3 projects like Chainlink are defining standards for KYC-bound, role-based agent identities.
In the meantime, retail and institutional platforms are shifting toward autonomous portfolios. Firms like Robinhood allow users to connect third-party AI agents that execute real-time trading strategies within defined guardrails. Prediction platforms like Polymarket deploy automated arbitrage bots to execute. Meanwhile, platforms like Polymarket, automated arbitrage bots execute lightning-fast contract trades to capture mispriced assets much faster than humans, keeping markets highly efficient.
Although AI regents are expanding exponentially in the tech sector, the machine economy faces severe growing pains that need to be resolved before achieving systemic stability. Attackers have successfully used malicious prompt injections to trick consumer AI models into executing unauthorized on-chain wallet transfers. The system is also plagued with ambiguous legal liabilities without a proper definition of who should be held legally liable for faults.
Additionally, blockchains need to support ultra-high-throughput, ultra-low-latency consensus layers so that network congestion never breaks a machine’s decision-making flow. That would address the execution loop bottleneck currently facing the system.
