Coinbase AI agent trading expands into US stocks, ETFs and micropayments

2 hours ago 28
Coinbase AI agent trading

Coinbase is betting that artificial intelligence agents are about to become real market participants, not just chatbots that answer questions. Starting this week, Coinbase AI agent trading now extends beyond crypto and derivatives into regulated US stocks and ETFs, letting autonomous software handle research, payments and trade execution inside a single account. The update, rolled out through Coinbase for Agents, pairs equities access with a new micropayments system called x402, which lets an agent pay small fees for data or model outputs the instant it needs them.

Key takeaways

  • Coinbase for Agents now supports trading US stocks and ETFs alongside the crypto and derivatives already available on the platform.
  • Securities trade through Coinbase Capital Markets Corp., a member of FINRA and SIPC, with Apex Fintech Solutions handling clearing and custody.
  • The platform offers thousands of US stocks and ETFs with 24/5 trading, fractional shares starting at $1, and zero commissions.
  • The x402 payments protocol has processed more than 230 million transactions and over $54 million in volume, with Coinbase acting as the top facilitator.
  • Users retain control through configurable spending limits, asset permissions and per-action approval guardrails.

Coinbase for Agents expands into US stocks and ETFs

Coinbase says its agent platform is turning into what it calls the “Everything Exchange” for AI-driven trading, and the addition of real equities is the clearest sign yet of that ambition. Agents that could already trade crypto and derivatives on Coinbase for Agents can now buy and sell fully owned US stocks and ETFs, the same securities available through Coinbase’s retail app.

Integration with regulated equities via Coinbase Capital Markets Corp.

The equities side of the launch runs through existing regulated infrastructure rather than a new, separate system. Securities trade through Coinbase Capital Markets Corp., a member of FINRA and SIPC, while Apex Fintech Solutions handles clearing and custody — the same backend that already supports Coinbase’s equities business. In effect, Coinbase opened up a regulated brokerage pipeline it already operated and pointed it at a new type of customer: automated agents acting on a person’s behalf.

Key features: 24/5 trading, fractional shares from $1, zero commissions

Once connected, an agent gets access to thousands of US-listed stocks and ETFs, 24/5 trading windows, fractional share purchases starting at $1, and zero-commission execution. Tokenized equities outside the US are also listed on Coinbase’s roadmap, though not yet available.

x402 payments protocol enables instant micropayments for agents

Trading well requires information, and information usually costs money — a problem that’s awkward for software that needs to move in seconds, not through monthly invoices. To solve this, Coinbase introduced x402, a payments protocol embedded directly within Coinbase for Agents, enabling an agent to pay a few cents for data or model inference—like an earnings summary or an onchain analytics query—and immediately act upon the outcome.

Real-time payments for data, models, and services within agent workflows

That’s the gap x402 is designed to close: built directly into Coinbase for Agents, the protocol lets agents pay each other in real time worldwide, even at fractions of a cent per transaction. Inside Coinbase for Agents, that same mechanism now lets an agent research a stock, pay for confirming data, and place the trade — all within one continuous loop and a single login.

High transaction volume and Coinbase as leading facilitator

The numbers suggest the protocol is already gaining real traction rather than sitting as a proof of concept. In just over a year, x402 has processed more than 230 million transactions and over $54 million in volume, and Coinbase says it is the number one facilitator, handling more than half of all x402 transactions industry-wide.

Why this matters for AI-driven markets

Pairing regulated equities access with programmable micropayments changes what an agent can realistically be asked to do. Instead of a person juggling brokerage logins, data subscriptions and API keys separately, one account now covers research, payment and execution. That collapses a workflow that used to require several tools into something closer to a single instruction: watch for a signal, pay to confirm it, act on it.

It also puts Coinbase in a position most trading platforms haven’t reached yet — offering a regulated bridge between AI agents and both crypto and traditional securities markets at the same time. If agentic trading keeps growing as expected, having that infrastructure already live, rather than in development, becomes a meaningful head start.

User control and unified account experience for AI agent trading

None of this hands an agent a blank check. Coinbase built configurable guardrails into the system: users decide which assets an agent can touch, how much it can spend, and whether every action needs individual approval. An agent can be locked into crypto-only trading, given a defined equities allowance, or restricted to specific tickers, and Coinbase says sandboxed accounts for equities are coming soon as an added layer of separation.

Configurable permissions, spending limits, and per-action approvals

Every trade and every data payment draws from the same account balance used for regular trading — there are no separate subscriptions, API keys, or accounts to fund and manage. That single-balance design is what lets an agent move between research, payment and execution without a person stepping in at each stage, while still keeping spending inside limits the user set in advance.

How Coinbase agents put the tools to work

Coinbase points to internal examples to show what the combination looks like in practice. One team member set an agent to research NVDA, AAPL and GOOG ahead of an earnings cycle; using x402, it paid a few cents for analyst sentiment and options-flow data, summarized the findings, and placed fractional buys across all three stocks. Another employee had an agent monitor an ETH position, watching onchain flow data for a spike in exchange inflows and automatically executing a preset stop-loss without anyone monitoring a screen. A third used an agent to track unusually active wallets each morning, paying a cent for the data, cross-referencing trending tokens, and buying a small position when predefined conditions were met — all within a set spending limit.

In each case, the person defined the strategy and the boundaries; the agent handled the watching, the paying and the executing.

Getting started and what comes next

Coinbase says setup takes as few as three clicks. Inside Perplexity or Grok, users can find Coinbase in the Connectors or Plugins tab and sign in to their account; ChatGPT, Claude and other agents can connect through MCP (model context protocol) directly. Tokenized equities beyond the US remain on the roadmap, suggesting the current US stocks-and-ETFs rollout is an early phase rather than the finished product.

Coinbase frames the move as an extension of its earlier push to make crypto accessible to a mass audience, this time applied to agentic trading. Whether autonomous agents end up handling meaningful trading volume — or simply automate the busywork around it — will likely depend on how comfortable users get with letting software spend on their behalf, even inside limits they control.

FAQ

What new asset classes can Coinbase for Agents trade?

Coinbase for Agents now supports trading of US stocks and ETFs in addition to crypto and derivatives.

How are securities handled and cleared on the platform?

Securities trade through Coinbase Capital Markets Corp., a FINRA/SIPC member, with Apex Fintech Solutions handling clearing and custody.

What is the x402 payments protocol and how does it work?

x402 is a payments protocol enabling AI agents to make instant micropayments for data, models, and other services, using the same account balance as trading.

How does user control over AI agents work on Coinbase for Agents?

Users can set agent permissions, spending limits, asset access, and per-action approval guardrails to control AI agent activities.

Article produced with the assistance of artificial intelligence and reviewed by the editorial team.

Read Entire Article