AI Agents Are Reshaping Financial Infrastructure as DeFi Trading Hits $68.5 Billion Monthly

AI Agents Drive Fundamental Shift in Transaction Infrastructure

AI agents are moving beyond content generation into autonomous economic activity, creating pressure on financial infrastructure that was never designed for machine-scale commerce1

1

. These autonomous AI agents now purchase computing power, access data, hire software services, and complete transactions without human approval for each step. Tom Lee, co-founder and head of research at Fundstrat Global Advisors, warns that if traditional payment infrastructure proves too slow or restrictive, machines could gravitate toward alternative systems for exchanging value1

1

.

The challenge stems from fundamental design differences. Current payment infrastructure reflects human needs—consumers make purchases, businesses pay suppliers, banks identify account holders. An autonomous system managing complex tasks may need to purchase small amounts of computing power, pay for individual API calls, or compensate other agents for services continuously and in volumes that make traditional payment processes impractical1

1

. Logan Xie, leader of KuCoin AI Lab, identifies the real gap as "not just speed, but a machine-readable framework for trust and authorization"1

1

.

DeFi Aggregators Route $68.5 Billion Monthly as Execution Layer

Decentralized finance has emerged as testing ground for AI-driven trading, with DeFi aggregators currently routing approximately $68.5 billion over the past 30 days according to DeFiLlama data2

2

. Jupiter alone routed about $14.1 billion over 30 days, while OKX DEX handled $7.29 billion, DFlow $6.53 billion, 0x roughly $6.27 billion, and KyberSwap $6.14 billion2

2

.

Major platforms including 1inch, LI.FI, KyberSwap, Uniswap, and Coinbase have launched infrastructure this year enabling AI agents to execute swaps, portfolio management, and onchain transactions2

2

. Rather than becoming exchanges themselves, AI agents in DeFi rely on aggregators as invisible execution infrastructure. LI.FI aggregates liquidity across 40-plus DEXs, aggregators, and bridges, while its infrastructure reaches more than 60 chains and over 1,000 applications2

2

.

Machine-to-Machine Commerce Demands New Payment Rails

Traditional card economics create fundamental barriers to machine commerce. Mark Zalan, CEO of GoMining, explains that "card economics put a floor of a few cents under every transaction, so a payment of a fifth of a cent simply cannot exist on those rails at any fee level"1

1

. The machine economy runs on exactly those microtransactions: compute, data, API calls, bought continuously in tiny increments1

1

.

Programmable payment networks operating around the clock can be accessed directly by software and use automated contracts to enforce transaction conditions. Rather than creating entirely new monetary systems, experts predict AI agents in DeFi will gravitate toward stablecoins, blockchains, or other programmable financial instruments1

1

. Zalan suggests machines "will gravitate to whatever settles fastest and cheapest with the fewest permissions, and the sum of billions of those cold, unsentimental choices will look, in retrospect, like a monetary order nobody voted for"1

1

.

Real-World Testing Reveals Performance Gaps

In March, LI.FI gave Claude Opus 4.6, GPT-5.4, Gemini 3.1 Pro, Grok 4.1 Fast, and MiniMax M2.5 $1,000 USDC each for seven days of autonomous trading across Ethereum, Arbitrum, Base, and Solana2

2

. Gemini finished first with a 5.57% gain while Claude finished down 11.75%. Claude made 38 trades, more than five per day, and agents that traded more frequently generally performed worse2

2

. The results show AI agents solved the mechanical problem of trading without solving the harder problem of knowing when to trade.

Ethereum Applications Shift to Intent-Based Architecture

AI agents and Ethereum applications are evolving toward intent-based interactions where users provide goals rather than manual steps3

3

. Instead of choosing tokens, selecting exchanges, approving spending, and executing swaps manually, an agent could receive instructions like "move $500 into the highest-yielding approved stablecoin strategy while keeping risk below a specified threshold"3

3

. Ethereum now has around $46 billion deployed across DeFi protocols according to Ethereum.org's May 2026 data, giving agents existing financial infrastructure they can call programmatically3

3

.

The x402 payment standard uses HTTP 402 Payment Required status to let software pay directly for APIs, data, and computing resources3

3

. Ethereum and Layer 2 networks support x402 stablecoin payments, allowing an agent to purchase single services without accounts or subscriptions. Coinbase expanded x402 in March to support almost any ERC-20 token through Permit2 and gas-sponsorship technology, enabling machine-to-machine commerce where one agent purchases data or computing power from another3

3

.

Infrastructure Providers Build for Autonomous Future

1inch launched Model Context Protocol integration in March, allowing agents to access swap infrastructure, portfolio data, and onchain services through standardized tools2

2

. Co-founder Sergej Kunz predicts agents rather than humans will execute the majority of swaps by 2030. KyberSwap launched its MCP in April with 13 composable tools covering trading, liquidity, limit orders, and Zap transactions2

2

. LI.FI launched an API specifically for agentic commerce in March, enabling agents to execute swaps, bridges, and multi-step DeFi transactions through a single interface2

2

.

Uniswap's open-source AI skills had surpassed 7,500 installations by July, when it added tools allowing coding agents to construct DCA strategies, token indexes, and automated rebalancing workflows from prompts2

2

. Smart accounts using Ethereum's account-abstraction infrastructure enforce spending limits, whitelists, session keys, and transaction permissions to restrict agents to specific applications or maximum transaction values3

3

. The Ethereum Foundation has already run coordinated AI agents against protocol code, identifying real vulnerabilities including a remotely triggerable libp2p gossipsub issue later disclosed as CVE-2026-342193

3

.

The investment question extends beyond chips, data centers, and models to systems that allow autonomous software to operate in the economy. Digital settlement assets and smart contracts provide machines with value that can move continuously through automated networks without relying on traditional banking processes for every transaction. Watch for continued concentration of machine economic activity around specific programmable payment networks and whether traditional financial institutions can adapt infrastructure fast enough to remain relevant for autonomous commerce.