The development prospects of AI combined with on-chain payments
Recently, Google released the Agent Payments Protocol (AP2), designed specifically for AI Agents.
Upon seeing this news, my first reaction was that the internet giant has finally extended its reach into the payment field for AI Agents.
Why do I say "finally"?
Because payments for AI Agents are not unfamiliar to users who have been following the crypto ecosystem over the past two years. The AI Agent BYTE in the Virtual ecosystem has long implemented payment applications and has already been integrated into Coinbase's application.
In the payment scenarios realized by BYTE, it uses the crypto asset USDC for transactions.
This formed my native impression of AI Agents payments: this scenario is inherently suitable for and should use crypto assets.
Therefore, I have always believed that the advancement in this field should be driven by native enterprises of the crypto ecosystem or at least by enterprises with crypto business.
Moreover, using crypto assets for payments means that all transactions are on-chain, which inherently possesses the most important core characteristics of payment business: transparency, verifiability, and immutability.
Since these characteristics already exist in the crypto ecosystem, while traditional internet giants do not possess them, what else can Google achieve by stepping in, apart from developing its own L1 for entering the financial business?
With this curiosity, I studied the content of its announcement (see reference link for details).
In this announcement, I believe the two typical scenarios it listed best reflect the purpose of releasing this protocol:
Real-time shopping (with humans present). A human instructs AI Agents to perform a purchase action and ultimately signs to allow payment. In this process, the human's instructions are recorded to prove that the tasks executed by AI Agents were indeed issued by a human, rather than being fraudulent.
Proxy tasks (without humans present). A human pre-arranges a purchase action that needs to be executed by AI Agents in their absence. In this process, the pre-arranged instructions are recorded to prove that the tasks executed by AI Agents were indeed arranged in advance by a human, triggered by certain conditions, rather than being fraudulent.
From the content presented in this announcement, my understanding of Google's AP2 protocol is that its core is not to reshape the process of on-chain payment transactions, but rather to prove that the payment transactions executed by AI Agents have solid reasons and sources, more accurately, that they come from humans.
Google's announcement listed several well-known partners in the Ethereum ecosystem, such as the Ethereum Foundation, EigenCloud (formerly EigenLayer), and Coinbase.
The Ethereum Foundation recently established an organization specifically targeting AI applications, particularly focusing on verification within AI applications. This indicates that the core team of Ethereum is beginning to take the AI track seriously, not wanting to be absent in this race that concerns the future. I wonder if this move is directly related to Google's protocol?
The original intention of EigenCloud/EigenLayer was to leverage the security of the Ethereum network to provide security services for other projects that also require security guarantees. However, after a few years, apart from a brief wave during its token launch, there has been little to no news about this project, let alone seeing any projects achieve significant results using its services. I hope AI Agents payments can bring real applications and scenarios to this project and its related track.
I believe Coinbase is more likely to directly benefit from this protocol.
The reason is simple: Coinbase's current application has already integrated several AI Agents that have implemented payment applications, and with the support of this protocol, its crypto payment closed loop will become even more perfect and secure.
On-chain payments were initially a scenario envisioned by AI + Crypto in the crypto ecosystem, and now traditional internet giants are also attempting to insert themselves into the AI + Crypto track.
The results explored by these two ecosystems from different angles surprisingly converge, which may indicate that in the near future, we will indeed see disruptive applications emerge in this scenario.
Popular articles














