Naval, a well-known Silicon Valley investor, posted a post a few days ago that has been read over a million times, saying: "Crypto is the currency of AI."

As if in response to his words, Coinbase CEO Brian Armstrong announced last Saturday that the first AI-to-AI crypto transaction had been completed on the Base network, with one AI using cryptocurrency (tokens) to purchase text content (tokens) generated by another AI.

Currently, AI cannot use traditional payment methods to purchase cloud computing services (such as AWS), developer tools (such as GitHub), or pay other expenses (such as booking flights or hotels). AI also cannot bypass paywalls to obtain the information it needs. These limitations greatly hinder the potential utility and actual work ability of AI.

The multi-party computation (MPC) wallet launched by Coinbase Developer Platform (CDP) can be integrated with AI Agent to enable it to make payments autonomously. Brian Armstrong encourages developers to integrate this wallet into their AI models to achieve payment capabilities, thereby helping AI obtain resources and services.

Coinbase did not just add payment functionality to AI Agent on a whim, but had been planning for a long time. As early as May this year, Brian Armstrong said that "self-hosted crypto wallets will provide support for AI Agent."

Subsequently, in early August, Coinbase not only personally made a demo product of AI paying cryptocurrencies to humans, but also launched a "CDP AI Builder Program" to provide funding for AI robot developers who build integrated Coinbase MPC wallets.

Why is Coinbase so keen to add payment functions to AI Agent? As Wang Chao, an OG in the DAO circle and an AI investor, said in an interview with BlockBeats, the first large-scale scenario of combining AI Agent with encryption is most likely that AI Agent uses encryption technology during settlement. The future Agent economy will form a huge emerging market, and the blockchain network that provides payment methods for Agents will also benefit from the huge amount of capital flow on it.

According to Wang Chao's estimate, AI technology is expected to create another 100 trillion in GDP in ten to fifteen years, and AI technology as the underlying support can obtain 20% of the share, that is, 20 trillion. Of this 20 trillion, electricity and computing power as the underlying support may take 10 trillion; the model layer, such as OpenAI, may take 5 trillion; the application layer takes 4 trillion; and the remaining service layer takes 1 trillion. Many products that create this 20 trillion will use encryption technology. Even if the combination of encryption and AI has only penetrated 1% of the overall AI market, this means that encrypted AI may get 200 billion from it. Therefore, Coinbase wants to become the "Visa/MasterCard in the field of AI" and seize this market opportunity worth 200 billion US dollars.

But Coinbase is not the only player in this field. Payman, invested by Visa, and Skyfire, invested by Circle, are both doing the same thing.

Coinbase Competitors

Payman

On August 7, Payman co-founder Tyllen announced on social media that it had completed a $3 million pre-seed round of financing, with participation from Visa, Coinbase Ventures, Boost VC, Deepwater Management, The Spartan Group, Untapped Capital, Theory Forge VC and others.

The idea of ​​Payman is to provide a platform API that allows AI Agents to pay humans in a variety of ways, including legal tender, cryptocurrency, and bank accounts. Before founding Payman, Tyllen worked as a product manager at Meta, and after entering the crypto industry, he worked as a developer relations director at Dapper Labs and Uniswap Foundation.

Although Payman provides more payment options for AI Agent, the platform is still in internal testing and there are not many details about how these functions are implemented.

Skyfire

On August 23, Skyfire, founded by former Ripple executives, announced the completion of a $8.5 million seed round of financing. Investors included USDC issuer Circle, Ripple, Gemini, and the venture capital firm of Tim Draper, a well-known Silicon Valley billionaire and Bitcoin advocate.

Skyfire has developed an open source payment system that allows AI Agents to conduct various transactions on the Internet, including the purchase of data storage, creative assets, air tickets, and daily necessities. Its payment network is based on the USDC stablecoin and is currently deployed on the Polygon network, with plans to expand to other blockchains in the future. For services that do not yet support USDC payments, Skyfire is considering integrating third-party aggregators to realize the transfer of USDC payments to traditional payment methods.

The above two projects, including Coinbase, can be regarded as a payment middleware platform, focusing on providing interfaces and services for payments between AI agents and humans. They are more targeted at AI developers rather than ordinary users.

However, there are currently some projects trying to make an AI Agent product with payment functions directly for ordinary users, such as Bitte, the AI ​​Agent wallet of the Near ecosystem, and Wayfinder, the AI ​​interaction protocol incubated by Parallel.

Bitte

Compared to Coinbase's idea of ​​integrating crypto wallet API into AI, Bitte chose to integrate the large model API into the crypto wallet. In the case of imperfect related infrastructure, Bitte tried to take a small step forward with existing technology. By accessing OpenAI's model API, Bitte allows users to use natural language prompts to command Agents to complete various on-chain operations in a chat window interface similar to ChatGPT.

For example, if the user enters "Mint an NFT with AI of a rocket going to the moon", Bitte will call the API of DALL-E 3 to generate an image of a rocket landing on the moon and mint it on the NEAR blockchain. In addition, users can also let the Agent complete a series of simple operations such as swapping and transferring tokens, creating contracts and NFT collections, etc.

Wayfinder

Wayfinder also allows users to use natural language prompts to command Agents to complete various on-chain operations in a chat window interface similar to ChatGPT. The protocol is still under development and is expected to be open to the public by the end of the year.

Compared with Bitte, Wayfinder wants to achieve more functions. Parallel co-founder Kalos said that theoretically its upper limit is the sum of all blockchain transactions, which means it can do anything. But he also admitted that Wayfinder currently lacks some underlying technical support, especially in the ability to allow AI to automatically plan and execute a series of blockchain operations according to user instructions (Prompt). To achieve such functions, Wayfinder needs to further develop or integrate the underlying protocol so that AI can handle complex tasks more intelligently.

Summarize

Although the combination of AI Agent and encryption is an inevitable trend of mutual benefit and win-win, it is still in a very early stage. Most of the projects mentioned above are in the proof-of-concept stage and there is still a long way to go before they can be truly applied. The biggest challenge is how to ensure the stability of large models in complex environments.

The immaturity of AI Agents means that most people cannot trust them with the management of their funds. Optimistically speaking, it may take several years before we see sufficiently powerful AI Agents emerge. In the future, more innovative projects will emerge that focus on providing payment functions for AI Agents.

Welcome to BlockBeats the BlockBeats official community:

Telegram subscription group: https://t.me/theblockbeats

Telegram group: https://t.me/BlockBeats_App

Official Twitter account: https://twitter.com/BlockBeatsAsia