‘AI agents can reason, but they cannot act’: MoonPay builds bridge to money

ambcryptoPublished on 2026-02-25Last updated on 2026-02-25

Abstract

MoonPay, a global leader in crypto payments and stablecoin infrastructure, has launched MoonPay Agents to enable autonomous AI systems to access and use capital. Recognizing that "AI agents can reason, but they cannot act economically without capital infrastructure," the company built a bridge between AI reasoning and real-world financial execution. This allows verified AI agents to programmatically trade, swap, and move digital assets via non-custodial wallets. This development aligns with broader infrastructure shifts, such as machine-to-machine payment protocols and a surge in stablecoin adoption. Stablecoins are transforming settlement with faster speeds and lower costs compared to traditional card networks, with transaction volume reaching $33 trillion in 2025. Institutional adoption is growing, with nearly half of institutions using stablecoins for payments. Despite regulatory scrutiny and bank resistance, the agent economy is expanding rapidly, compressing legacy financial margins and reshaping the future of autonomous commerce.

AI agents and stablecoins have reached a structural milestone in early 2026. MoonPay has recognized that shift and moved before hesitation set in.

Founded in 2019, MoonPay has reached more than 30 million customers across 180 countries. It supported over 500 enterprise clients. Through one integration, it powered on and off ramps, trading, crypto payments, and stablecoin infrastructure.

Therefore, MoonPay Agents did not emerge from a speculative lab. It came from a scaled operator already embedded in global digital asset flows. That context mattered.

MoonPay agents and financialization of autonomous AI

MoonPay, the leader in global crypto payments and stablecoin infrastructure, formalized capital access for programmable systems.

Its software layer enabled non-custodial wallets through MoonPay CLI. Once verified, agents traded, swapped, and moved digital assets programmatically. CEO Ivan Soto-Wright stated,

“AI agents can reason, but they cannot act economically without capital infrastructure.”

That line was blunt. Intelligence without money meant nothing. MoonPay built the bridge between reasoning and execution.

Infrastructure layer powering machine-to-machine payments

The launch aligned with deeper infrastructure shifts.

x402 enabled wallet authenticated HTTP payments using USDC without API keys. Quicknode supported x402 across more than 80 chains. Meanwhile, Coinbase’s Payments MCP recorded a 10,000% spike in agent transactions on Base.

As a result, machine-driven settlement was already accelerating. The rails existed. MoonPay plugged directly into them.

How stablecoins are reshaping settlement in AI era

Stablecoins are rewriting settlement math.

Card networks charged 2% to 3.5% per transaction. Cross-border flows exceeded 4%. Stablecoin transfers are settled within seconds at fractions of a cent. Remittance fees still averaged 6.6%.

Stablecoin transaction volume reached $33 trillion in 2025, rising 72% year over year. Supply surpassed $300 billion.

McKinsey & Company, a global consulting firm, estimated real world stablecoin payments approached $390 billion annually. Citi projected supply could reach $1.9 trillion by 2030.

Institutional adoption and expanding role of agentic commerce

Institutional readiness was no longer theoretical.

Fireblocks research showed nearly half of institutions used stablecoins for payments. Over 80% reported infrastructure readiness. Agent tokens expanded rapidly, with 21,000 minted in November 2024 alone.

Capital markets reacted sharply. Visa fell 4.6%, Mastercard dropped 5.7%, and American Express slid 7.2%.

According to a thesis published by Bull Theory, AI selecting the cheapest settlement rails could compress legacy margins.

What are execution and regulatory risks in agent economy?

Despite momentum, friction remained real.

Banks resisted yield-bearing stablecoins due to deposit migration concerns. Regulators intensified scrutiny around reserve backing and compliance.

However, incumbents quietly integrated stablecoin settlement infrastructure. They were not surrendering. They were adapting.


Final Summary

  • MoonPay launched MoonPay Agents, enabling autonomous AI systems to access capital through non-custodial wallets and programmable infrastructure.
  • The company leveraged its scale of 30 million users and 500+ enterprise clients to integrate AI into real crypto payment rails.

Related Questions

QWhat is the core problem that MoonPay Agents aims to solve for AI, according to CEO Ivan Soto-Wright?

AThe core problem is that while AI agents can reason, they cannot act economically without capital infrastructure. MoonPay Agents provides the bridge between reasoning and execution by enabling access to capital.

QWhat key infrastructure shift, mentioned alongside x402 and Quicknode, is enabling machine-to-machine payments?

AThe key infrastructure shift is the enablement of wallet authenticated HTTP payments using USDC without API keys, supported by x402 and Quicknode across more than 80 chains.

QHow do the transaction costs of stablecoins compare to traditional card networks and remittance fees?

AStablecoin transfers are settled within seconds for fractions of a cent, which is significantly cheaper than card networks that charge 2% to 3.5% per transaction and remittance fees that average 6.6%.

QWhat was the market reaction of traditional payment companies like Visa and Mastercard to the rise of stablecoins and agentic commerce?

ACapital markets reacted sharply, with Visa falling 4.6%, Mastercard dropping 5.7%, and American Express sliding 7.2%, as AI selecting cheaper settlement rails threatened to compress legacy margins.

QWhat are two major sources of friction or risk mentioned for the growth of the agent economy?

AThe two major sources of friction are banks resisting yield-bearing stablecoins due to deposit migration concerns, and regulators intensifying scrutiny around reserve backing and compliance.

Related Reads

Bitcoin Trading Strategy Breakdown: Celebrity Predictions and Classic Models All Fail, Only These Four Indicators Remain

Analysis of Bitcoin Trading Strategies: Why Celebrity Forecasts and Classic Models Fail, Leaving Only These Four Reliable Indicators This analysis examines the failure of common Bitcoin prediction methods and identifies four reliable indicators for constructing a trading strategy. The author reviewed all major BTC prediction approaches from 2017-2025, categorizing them into three groups: celebrity price targets (consistently over-optimistic), analytical models like Stock-to-Flow (broken post-2022), and on-chain signals. The key finding is that more data often creates confusion, not clarity. The strategy discards unreliable elements: celebrity predictions (incentivized to be extreme), pure models (invalidated by post-ETF market changes), and the Fear & Greed Index used alone (too many false signals). Four reliable indicators were selected: 1. **MVRV Z-Score:** Accurately identifies cycle bottoms when entering its green zone (e.g., 2018, 2020, 2022). Note: Its ability to call tops is now ineffective post-2024. 2. **SOPR (28-day MA):** Consistently signals bottoms when below 1.0, indicating holders are selling at a loss. 3. **ETF Net Flow:** A crucial post-2024 metric showing institutional momentum (e.g., sustained inflows = buying). 4. **Macro Liquidity (Fed policy & M2):** Sets the overall directional bias (e.g., bullish during easing cycles). The core strategy involves waiting for a multi-signal共振 (resonance). For example, a bottom signal requires MVRV in the green zone + SOPR < 1.0. A top signal requires overheated on-chain data + sustained ETF outflows. Macro policy sets the overall direction. The Fear & Greed Index is only used as a weighted confirmatory signal, never alone. Action is only taken when three or more indicators align. The author automated this into a monitoring system that sends Telegram alerts only when signals trigger. As of the article's date (April 15, 2026), the system showed a strong bottom signal: extreme fear (F&G=12), MVRV in the buy zone, and SOPR < 1.0. The only contrary signal was weak ETF flows. Historically, such triple on-chain共振 has preceded 100%+ returns. The conclusion emphasizes building a personal framework over relying on external predictions, allowing for iterative improvement and customization based on individual risk tolerance.

marsbit1h ago

Bitcoin Trading Strategy Breakdown: Celebrity Predictions and Classic Models All Fail, Only These Four Indicators Remain

marsbit1h ago

Trading

Spot
Futures

Hot Articles

Discussions

Welcome to the HTX Community. Here, you can stay informed about the latest platform developments and gain access to professional market insights. Users' opinions on the price of AI (AI) are presented below.

活动图片