AI Agents Could Become Crypto's Most Active Users

For fifteen years, the crypto industry has built lending protocols, stablecoin rails, and tokenization infrastructure largely designed for human users. A provocative new thesis suggests that the most active consumers of that infrastructure may not be people at all — they may be AI agents.

Investor and analyst Jordi Visser outlined this vision in what he calls the "Ghost Rails" concept: a future where autonomous AI agents, operating around the clock without human intervention, become the primary drivers of payments, asset transfers, and settlement on blockchain networks.

The Ghost Rails Thesis

The name captures the idea elegantly. "Rails" refers to the financial infrastructure — payment networks, settlement layers, lending markets — that move value between parties. "Ghost" refers to the agents themselves: software entities with no physical presence, no bank accounts in the traditional sense, and no need for sleep.

Visser argues that crypto's existing infrastructure is uniquely suited to serve these agents. Traditional banking systems require human identity verification, business hours, and manual approval workflows. Blockchain networks operate 24/7, settle transactions in seconds or minutes, and accept programmatic interaction through smart contracts.

An AI agent managing a portfolio, executing trades, or settling payments between services does not need a checking account. It needs a wallet, a stablecoin balance, and access to DeFi protocols. Crypto provides all three.

Why Agents Need Crypto, Not Banks

Consider what an autonomous agent actually does in practice. It receives a task — monitor a supply chain, manage a treasury, execute a trading strategy — and then takes actions to complete that task. Those actions often involve moving money.

In the traditional financial system, every transaction requires human authorization at some point. An agent cannot open a bank account. It cannot sign a wire transfer form. It cannot call a broker to place an order during off-hours.

Crypto removes these frictions. A smart contract can hold funds and release them when predefined conditions are met. A stablecoin transfer settles in seconds regardless of time zone. A DeFi lending protocol can extend credit based on collateral rather than credit scores.

For agents operating at machine speed, these properties are not conveniences — they are requirements.

Tokenization and the $900 Trillion Opportunity

Visser also points to tokenization as a complementary force. He estimates that roughly $900 trillion in global assets — real estate, private equity, art, intellectual property — remains illiquid and difficult to trade.

Tokenization technology can convert these assets into digital representations that move on blockchain networks. If AI agents become the primary traders and managers of tokenized assets, the demand for crypto infrastructure expands dramatically beyond its current role as an investment vehicle.

Imagine an agent that monitors real estate markets, identifies undervalued properties, tokenizes fractional ownership, and distributes returns to stakeholders — all without a human touching the process. That scenario requires lending markets for leverage, stablecoins for settlement, and tokenization platforms for asset representation. The crypto industry has been building exactly these tools.

Bitcoin as the Reserve Asset for Agents

Visser is particularly bullish on Bitcoin in this framework. As the agent economy grows, he argues, demand will concentrate on assets that have proven durable, widely accepted, and resistant to manipulation. Bitcoin's fixed supply, global liquidity, and decade-plus track record position it as a natural reserve asset for agent treasuries.

This is not the same as saying Bitcoin will replace the dollar for human commerce. It is a narrower claim: that autonomous systems managing digital value will prefer neutral, programmatically accessible assets over fiat currencies tied to banking systems designed for humans.

The Infrastructure Is Already Built

What makes Visser's thesis compelling is timing. The lending protocols (Aave, Compound), stablecoin networks (USDC, USDT), and tokenization platforms (Centrifuge, Maple) already exist. They are not theoretical. They process billions of dollars in volume.

The missing piece has been demand from non-human actors operating at scale. AI agents may supply that demand. Every major AI lab is investing in agent capabilities — systems that can browse the web, write code, manage files, and interact with APIs autonomously. Extending those capabilities to financial infrastructure is a natural next step.

Risks and Counterarguments

Not everyone agrees this is positive for crypto. Phemex co-founder Federico Variola has argued that AI is a "net negative" for the industry, draining liquidity while making it easier for attackers to discover protocol vulnerabilities. The July 2026 Coldcard wallet attack, which stole approximately $116 million in Bitcoin from over 5,200 addresses using suspected AI-assisted methods, illustrates the downside.

Regulatory uncertainty also looms. If agents begin moving significant value on-chain, governments will scrutinize the infrastructure enabling those flows. The CLARITY Act's recent stall in the U.S. Senate shows that regulatory clarity remains elusive.

What to Watch

The Ghost Rails thesis will be tested in the coming months as AI agents move from research labs into production deployments. Watch for:

Stablecoin transaction volumes driven by programmatic wallets rather than human users
DeFi protocols launching agent-specific APIs and permission systems
Major AI labs integrating crypto payment rails into agent frameworks
Regulatory responses to autonomous financial activity on public blockchains

Whether Visser is right that agents become crypto's dominant users or wrong that the timeline is imminent, the intersection of AI and blockchain infrastructure is one of the most consequential technology stories of 2026. The rails are built. The ghosts may already be arriving.