Crypto’s promise of mass adoption has echoed for more than a decade but despite 2026 now coming to a close, it is clear that the wait is still on. To this point, even though about 30% of American adults own some form of crypto, the numbers are still at the same level they were back in 2023.

Now, if one were to compare these same figures with AI, the trend seems quite clear, i.e., ChatGPT alone reached 900 million weekly users by Q1 of this year (with over five percent of these individuals being paying customers). Similarly, Stanford's latest AI Index has revealed that generative AI is already seeing mass adoption levels within just three years of the tech going mainstream.

Put simply, while one technology (i.e. crypto) is riddled with complexities such as seed phrases, gas fees, and network selection, the other just asks users to type in a question.

The friction was never just volatility

From the outside looking in, price swings still top the list of worries when it comes to crypto followed by a long list of usability complaints, such as:

  • Fear of cyberattacks
  • Distrust of exchanges
  • Risk of losing access to one’s funds

These worries are not irrational. In its 2025 Internet Crime Report, released this April, the FBI recorded approximately $11.4 billion in reported losses involving cryptocurrency, with people aged 60 and over accounting for roughly $4.4 billion.

AI could help reduce some of the operational complexity users face. Coinbase’s Agentic Wallets, for instance, isolate private keys from the AI agent and provide programmable spending limits. The product is aimed at developers, but the design points toward a broader possibility.

Keeping private keys out of an agent’s reach is an important safeguard, but it does not guarantee that every transaction it requests is appropriate. Spending limits need to be enforced outside the model, permissions should be easy to revoke, and users need a clear record of what they authorized and what was executed. As the interface becomes simpler, those controls become more important.

Picture someone asking a chatbot to send $50 to a relative abroad. A well-designed system could identify a supported payment route, calculate the total fee and show what the recipient would receive before asking for confirmation. Stablecoins could provide the settlement layer where they offer an advantage in cost, speed or availability.

Users would not need to select the network manually, but they would still need to understand and approve the payment. The recipient would also need a practical way to use or withdraw the funds.

The crypto side of the transaction could become largely unseen infrastructure. The meaningful financial decisions would remain visible.

The agent economy is still early

The infrastructure for this shift is already taking shape. Coinbase’s x402 payment protocol, launched in May 2025, enables stablecoin payments over HTTP, allowing applications and agents to pay for resources such as APIs and digital services.

This extends the opportunity beyond consumer payments. An assistant completing a task could purchase access to the information or computing resources it needs, within a budget set by its user.

Still, transaction counts alone cannot establish how much useful commerce is taking place. The more meaningful questions concern what is being purchased, whether the service is delivered and whether people return because the experience works.

Security requires similar care in interpreting the numbers. The FBI’s recent report included its first dedicated section on AI, recording 22,364 AI-related complaints and nearly $893 million in reported losses. These figures cover AI-related crime broadly, including impersonation and scams, rather than losses specifically caused by agentic wallets.

For AI systems handling payments, the implications are clear: hard spending limits, understandable confirmations and clear responsibility when something goes wrong. A record showing that money moved is only part of the picture. Users also need to know whether the action matched their instructions.

Where the next wave probably comes from

The next wave of adoption may come from people who want a simpler way to complete a task, whether that means sending money or paying for a digital service. AI assistants could give them a familiar interface, with crypto providing the underlying infrastructure where it offers a practical advantage.

The opportunity is to make those services useful without requiring people to become crypto experts first.

None of this can happen on its own. If AI makes crypto easier to use without making it safer, it risks making scams and mistakes more efficient. But if the guardrails hold up, the next jump in adoption may look less like a rush to sign up for exchanges and more like people moving money through AI assistants, with crypto running quietly underneath.