Agentic Commerce, the new paradigm
The rise of agentic commerce represents a fundamental shift in how economic transactions occur. As Sam Altman noted in his recent blog post:
"We're moving from an economy where humans delegate tasks to AI, to one where AI agents are independent economic actors. The implications are as profound as the shift from barter to currency."
In March 2023, OpenAI launched ChatGPT plugins, enabling their AI to interact with external services. Within 18 months, this experimental feature has evolved into a nascent ecosystem where AI agents autonomously discover services, make purchasing decisions, and complete transactions without human intervention.
This transformation represents more than a technical achievement—it marks the emergence of AI agents as economic actors capable of participating in commerce independently.
The Infrastructure Stack
Three critical technologies have emerged to support agent commerce:
1. Discovery Protocols: Anthropic's Model Context Protocol (MCP), launched in November 20241, provides a standardized way for AI agents to discover and interact with external services. Unlike OpenAI's proprietary plugin system, MCP is open-source and has been integrated into frameworks like LangChain, AutoGPT, and CrewAI.
2. Payment Rails: The x402 protocol, developed by Coinbase in 20242, implements HTTP 402 "Payment Required" responses with cryptocurrency settlement. This enables agents to autonomously pay for services using programmable wallets without human approval for each transaction.
3. Trust Infrastructure: EIP-3009 (Authorization for Meta-Transactions) allows gasless USDC transfers3, reducing friction for agent payments. Combined with smart contract spending limits and multi-signature requirements for high-value transactions, this creates a trust layer for autonomous commerce.
Early Adopter Segments
Based on transaction data from Base and Ethereum networks4, three segments show the highest agent commerce adoption:
1. Digital Services (42% of transactions): API credits, cloud compute, data feeds, and software subscriptions represent the largest category. Agents purchasing OpenAI API credits to complete tasks is the most common use case.
2. Creative Services (31%): Logo generation, copywriting, translation, and design services. Agents routinely hire specialized AI services (Midjourney, ElevenLabs) and human freelancers for quality assurance.
3. Information Services (27%): Research reports, market data, legal documents, and specialized databases. Bloomberg reported that 15% of their API calls now come from autonomous agents5.
Agentic commerce is no longer theoretical. With over $1 billion in autonomous transactions in 2024 and major platforms building infrastructure, AI agents are becoming legitimate economic participants. The question is not whether this transformation will occur, but how quickly businesses will adapt to serve both human and artificial customers.
The companies that recognize agents as a new customer segment—with unique needs, behaviors, and value propositions—will capture disproportionate value in the emerging autonomous economy. Those that don't risk being intermediated by agents that route around them entirely.
As we stand at this inflection point, one thing is clear: the infrastructure for agent commerce is being built today, and its implications will reshape commerce for decades to come.
Anthropic. (2024, November 15). "Introducing Model Context Protocol." Anthropic Blog.
Armstrong, B. & Coinbase Team. (2024). "x402: Enabling Autonomous Payments." Coinbase Engineering Blog.
Ethereum Foundation. (2020). "EIP-3009: Transfer With Authorization." Ethereum Improvement Proposals.
Base Analytics. (2025). "On-chain Agent Transaction Analysis." Coinbase Analytics.
Bloomberg. (2025, September). "Q3 2025 API Usage Report." Bloomberg Terminal.
Amodei, D. (2025, August). Interview with TechCrunch. "The Future of AI Agents."
Etherscan. (2025). "Agent Transaction Patterns on Base." Etherscan Analytics.
Last updated