Coinbase has unveiled plans to build financial accounts tailored for artificial intelligence (AI) systems and autonomous agents, aiming to expand its agentic finance initiative. CEO Brian Armstrong stated on X that the company is creating a financial account designed for the needs of AI, extending work already underway with agent wallets and stablecoin payment tools that support software-based trading and transactions.
Coinbase ramps up agentic wallet infrastructure
Coinbase has been working to integrate digital asset management solutions for AI agents. Its Agentic Wallets, introduced in February, enable autonomous spending, earning, and trading, while incorporating security safeguards. These wallets allow software agents to make payments directly, addressing a key limitation where agents previously could only recommend actions rather than execute them independently.
Building on the infrastructure for AI systems, Coinbase for Agents was launched in June. This service lets approved AI agents connect to Coinbase user accounts and perform trades or make payments within limits set by human users. The coming expansion will see these capabilities scale up from wallets to broader AI-oriented financial accounts, following Coinbase’s ongoing product development strategy rather than introducing a completely new model.
Coinbase is building the financial account for AI, aiming to streamline how autonomous agents manage assets, make payments, and execute trades under secure, user-defined conditions.
Rise of AI agents in global payments
Autonomous agents increasingly require the ability to purchase computing power, access APIs, and pay for various digital services directly. To address this, Coinbase developed the x402 protocol, which supports machine-to-machine payments in stablecoins like USDC. This infrastructure allows agents to transact with digital services seamlessly, reducing the need for human intervention in routine purchases.
In May, Coinbase reported processing over 100 million payments through the x402 protocol, with more than 99% settled in USDC. While not all these transactions involved AI agents, the volume underscores the scalability and adoption of its machine-oriented payments platform.
Mini dictionary: x402 protocol, a Coinbase-developed payments layer allowing autonomous machine-to-machine stablecoin transfers and designed to serve the transactional needs of digital and AI agents.
| Payment protocol | Total payments | Main stablecoin used |
|---|---|---|
| x402 | 100 million+ | USDC (99%+) |
Stablecoins power autonomous agent transactions
Stablecoins have played a central role in Coinbase’s approach, supporting dollar-denominated transfers across blockchain networks without relying on conventional banking or card systems. Coinbase has integrated USDC and other stablecoins within its payment platform, while its Base network serves as an Ethereum-layer-2 solution for processing these transactions efficiently.
The payments stack now supports both agent-initiated and enterprise-focused payment flows, aiming to make programmable payments accessible for software-driven activity. Expanding these capabilities enables a broader range of financial activity by autonomous agents across digital platforms.
Agentic finance becomes wider strategy for Coinbase
Coinbase is advancing its agentic finance model across trading, payments, and comprehensive financial management. Its broader vision connects autonomous agents with user accounts and introduces tools like Advisor, an AI-powered interface for retail users. Coinbase has also upgraded its business payments system to enable AI-initiated transactions at scale.
Financially, Coinbase reported a third consecutive record for quarterly crypto trading volume market share in the second quarter, reflecting a push to diversify products beyond traditional trading in response to evolving market dynamics.
Shares of Coinbase closed at $190.02 on September 29, slipping 0.92% that day. However, this decline cannot be directly linked to the agentic finance developments.
As Coinbase advances its agent-focused roadmap, the main challenge will be developing a scalable, secure financial framework for software agents. Issues such as identity verification, spending control, liability, fraud prevention, and regulatory compliance are expected to take on greater significance as autonomous agents gain more financial independence.
The company’s expansion moves from agent-specific wallets toward comprehensive accounts, aiming to form the backbone of non-human economic activity in digital finance.
