Agentic artificial intelligence (AI) is becoming a key player in the world of foundation models. When we say "agentic," we mean AI systems that can not only create outputs but also take actions in their environment. Think about tasks like replying to emails, interacting with databases, calling APIs, or even performing physical tasks with robots.

Unlike older automation models, agentic AI can allow agents to operate with some level of autonomy. Imagine a future where millions of these semi-autonomous agents play vital roles in both business and personal life. In this scenario, crypto assets could serve as the main currency that these AI agents use to interact with the world.

This raises an interesting idea: cryptocurrency might become the currency of AI. However, while this concept is exciting, there are some practical challenges that we need to consider.

Many Demos, Few Products

To really understand how AI agents can use crypto, we should look at the current state of the agent market. The excitement around semi-autonomous agents has led to a lot of new frameworks for building them. But the market isn’t quite ready yet. Most organizations see the potential of these agents but are still figuring out how to adopt them. This results in many demos but very few products that are ready for real-world use.

This is important because the success of combining crypto with autonomous agents depends on how well the agent market develops. If we envision a world where millions of these agents help with our daily tasks, crypto could enhance their capabilities significantly. There are many possibilities here, especially in the financial sector.

When we think of autonomous agents as extensions of our business processes, e-commerce becomes a crucial part of how humans and agents interact with each other. Cryptocurrency offers a promising path to unlock this potential.

The Economic Layer of the Intelligent Internet

Payments have always been central to the internet's evolution. PayPal opened up many early e-commerce opportunities, and platforms like Stripe have taken that transformation even further. If AI agents become the backbone of the intelligent internet, financial transactions will be key to this new layer.

The programmability of crypto makes it a great fit for enabling payments in AI agents. But what specific situations make crypto ideal for these workflows while showing the limitations of traditional fiat currency?

Imagine a scenario where one agent hires another to handle a specific task, like paying invoices and following up with clients via email. In this case, agents might receive many tasks, with payments released as they complete each one. However, the current state of the AI agent market poses challenges for making these scenarios work in real life. So, the path to adopting crypto for these transactions might not be straightforward and could take several iterations.

Cultural Clash Between Crypto and AI

Another barrier to adopting crypto in multi-agent systems is the cultural gap between the AI and crypto communities. Many teams developing agent platforms don't have strong connections to the crypto market. They may not be enthusiastic about Web3 and often see crypto as a hurdle to broader platform adoption. This disconnect also affects investors.

Many leading AI agent platforms got funding during tough times in the crypto market when venture capitalists were skeptical about its value. Additionally, organizations adopting AI agent platforms may still see crypto as a barrier to their operations.

While strong AI teams are emerging in the Web3 and AI agent space, this cultural friction shouldn’t be underestimated. Overcoming these challenges will require finding use cases that can’t be achieved without crypto payments.

Lack of Agent E-Commerce

In the crypto world, people often discuss the inevitability of AI agents conducting financial transactions. While the promise is clear, the necessary infrastructure is still lacking. Companies like Stripe built their businesses on years of e-commerce innovation, backed by extensive merchant networks. This kind of infrastructure isn’t yet in place for AI agents. As a result, early use cases for agents transacting may resemble the internet of the 1990s and will need significant evolution.

Let’s explore a hypothetical scenario involving AI agent transactions. Imagine a group of agents assigned 1,000 tasks across different locations, with payments unlocked upon completing each task. To finish these tasks, some agents may need to collaborate. While some agents complete their tasks quickly, others might struggle, leading to various levels of conflict.

We currently have three options in the market:

  1. Programmable Fiat: Platforms like Stripe offer a low barrier to entry for financial transactions in multi-agent systems. Stripe’s recent launch of an agent toolkit, integrated with many top frameworks, highlights this potential. The wide reach of programmable fiat platforms, their large developer communities, and seamless banking integration make them an attractive starting point for agent payments. However, there are limitations, such as challenges with cross-border payments and lengthy dispute resolution processes.
  2. Stablecoins: Stablecoins sit at a natural intersection between AI and Web3. They address some of fiat's limitations in agent workflows by offering on-chain settlements and global availability. Agents could receive payments in stablecoins like USDC or USDT upon task completion. However, stablecoins mainly serve as payment tokens and lack functionalities tied to agent performance, making dispute resolution outside the stablecoin system necessary when tasks aren’t completed.
  3. A Native Agent Cryptocurrency and Payment Protocol: The true intersection of Web3 and agentic AI for financial transactions could emerge through new tokens and protocols tailored for this use case. These could extend the capabilities of stablecoins by integrating agent-specific functionalities. In our scenario, payments could be made using a specialized asset that agents stake for quality control. Slashing policies could penalize poor performance, while validators could resolve disputes based on task quality.

Agents Are All We Need

Agentic AI shows great potential as a key intersection between Web3 and AI. Semi-autonomous agents are expected to become commonplace, requiring robust financial infrastructure for transactions. While programmable fiat may address initial needs, its limitations highlight the potential for cryptocurrency to excel.

Stablecoins offer a promising start, but they may not be enough in the long run. The rise of agentic AI is likely to drive the creation of custom tokens, wallets, protocols, and possibly blockchains optimized for agent workflows. Agents could become the gateway for crypto to truly emerge as the currency of AI.