Yesterday, Coinbase CEO Brian Armstrong posted something that stopped the crypto-AI world in its tracks:
"Very soon, there are going to be more AI agents than humans making transactions. They can't open a bank account, but they can own a crypto wallet. Think about it."
Within hours, Binance founder CZ amplified the signal: "AI agents will make 1 million times more payments than humans, and they will use crypto."
These aren't speculative takes from newsletter writers. These are the CEOs of the two largest crypto exchanges in the world, saying the same thing on the same day. And they're not just talking — Coinbase launched Agentic Wallets in February 2026, built on the x402 protocol, and has already processed over 50 million transactions.
The agent economy isn't coming. It's here. But there's a gap nobody is talking about.
The Problem Nobody's Solving
Armstrong nailed the identity problem: agents can't KYC into a bank, but they can hold a crypto wallet. That's the access layer — solved.
But what happens when those agents actually need to work?
Think about what a real AI agent workflow looks like. An orchestrator agent gets a task. It breaks that task into subtasks. It hires five specialist agents — one for data retrieval, one for analysis, one for formatting, one for quality checks, one for delivery. When the job is done, the orchestrator needs to pay all five agents in a single transaction.
Now multiply that by a million concurrent workflows.
Individual wallet-to-wallet transfers don't scale for this. You'd be burning gas on five separate transactions every time an agent completes a job. At the volume Armstrong and CZ are describing — agents making a million times more payments than humans — that's not just inefficient. It's broken infrastructure.
What's missing is the settlement layer. Batch payments. Multi-party programmatic payouts. The ability for one agent to pay 50 others in a single on-chain transaction.
That's what we built.
💧 Spraay: Batch Payment Infrastructure for the Agent Economy
Spraay is a multi-chain batch crypto payment protocol. One transaction, multiple recipients, one gas fee. Live on 11 chains including Base, Ethereum, Arbitrum, Polygon, Solana, and more.
But here's where it gets interesting for the agent economy specifically.
The x402 Gateway
We built an x402-powered API gateway at gateway.spraay.app — the same x402 protocol that Coinbase's Agentic Wallets run on. It exposes 57 paid endpoints across 13 categories that agents can call and pay for with micropayments:
- Batch payments — pay multiple wallets in one transaction
- RPC access — multi-chain node access across 7 networks via Alchemy
- AI inference — on-demand model calls through OpenRouter
- Messaging — XMTP agent-to-agent communication
- Data oracles — real-time token prices and on-chain data
- IPFS storage — decentralized file pinning via Pinata
- Email — agent-native email through AgentMail
- Escrow, payroll, and DeFi automation
Every endpoint is pay-per-call. No API keys. No subscriptions. No KYC. An agent with a wallet can start using Spraay infrastructure immediately — exactly the way Armstrong describes the future should work.
The MCP Server
For AI agents built on frameworks like Claude, we published a Model Context Protocol (MCP) server on Smithery with 60 capabilities. This means any MCP-compatible agent can discover and use Spraay's payment tools natively, without custom integration work.
Agent Commerce Protocol (ACP)
We're registered on the Virtuals Protocol ACP with structured job offerings — so agents on that network can hire Spraay's agent to handle their payment operations. Three revenue layers stack: x402 micropayments for API calls, ACP job fees for task completion, and aGDP token rewards.
Why Batch Payments Are the Missing Primitive
Let me paint the picture more concretely.
Scenario: Agent Payroll
A DAO runs 200 contributor agents. Monthly payroll day. Without batch payments, that's 200 separate transactions, 200 gas fees, 200 confirmations to track. With Spraay, it's one transaction. Done.
Scenario: Agent Marketplace Settlement
An AI agent marketplace processes 1,000 completed jobs per hour. Each job involves paying the service agent and taking a platform fee. Without batch: 2,000 transactions per hour. With Spraay: batched settlements every block, fractional gas costs.
Scenario: Multi-Agent Collaboration
A research agent orchestrates 10 specialist agents to produce a report. It collects payment from the client, then distributes shares to all contributors proportionally. One Spraay batch call handles the entire split.
This isn't theoretical. The smart contracts are deployed and verified. The gateway is live. The MCP server is published.
The Infrastructure Race Is On
Armstrong and CZ are telling us the demand side is coming — millions of agents needing to transact. Coinbase is building the wallet layer. What's needed now is the payment operations layer — the infrastructure that lets those wallets do complex, multi-party, programmable payments at machine speed.
That's the lane we've been building in for months, and we're already live on 11 chains.
If you're building AI agents that need to handle payments — whether that's paying for compute, splitting revenue, settling marketplace transactions, or running automated payroll — Spraay is the batch payment primitive you're looking for.
Links:
💧 Spraay — Batch crypto payments for the agent economy.
Top comments (0)