header-langage
简体中文
繁體中文
English
Tiếng Việt
한국어
日本語
ภาษาไทย
Türkçe
Scan to Download the APP

When AI Becomes the "User": OKX Chief Growth Officer Lennix Discusses Onchain OS and the Next-Generation On-Chain Transaction Paradigm

Read this article in 11 Minutes
Future Agent-Driven Transactions and Interactions Could Surpass Human-Led Markets
Source: OKX


The annual Hong Kong Web3 Carnival has once again kicked off, bringing together global crypto enthusiasts at the summit.


As one of the largest and most influential Web3 events in the industry, the "2026 Hong Kong Web3 Carnival" has gathered many heavyweight speakers and over a hundred exhibitors, attracting widespread attention and participation from developers, institutions, and industry participants worldwide.


On the first day of the summit, OKX Global Chief Business Officer Lennix was invited to attend and delivered a keynote speech titled "Start building on Onchain OS" at the main forum. The following is the content of the speech (edited for clarity):



Today, I want to talk about the future and the role of AI in our entire financial system, especially in our industry.


AI Is Not Just a Tool Upgrade, But the Beginning of a Trend


In the past year to a year and a half, changes have actually been very rapid. Everyone is talking about AI and using agents. With the development of large models, we are beginning to see agents not just as a concept but actually entering our business. We see users creating all kinds of agents to achieve different goals. Some people use OKX Wallet data, combine it with AI models to monitor the market, generate signals, and then deploy different strategies.


These applications have made us realize one thing: this is not just a tool upgrade but an emerging significant trend. Moreover, this trend may be very important because it will change the very nature of trading behavior and also how we interact on-chain.


OKX Wallet has spent a lot of time optimizing user experience, from Web2 to Web3, from interface to transaction flow. However, if agents become part of users in the future, even the main users, the entire logic of interaction will need to be redesigned.


So the current problem is actually clear. You can use agents for in-depth data analysis, and AI can generate many trading signals. But the key step, or what agents currently lack the most, is how to execute and how to execute securely. How can you maintain control while not sacrificing the autonomy of the agent? These are the real challenges.


We are now seeing that some Agents are able to do certain things, such as account management, initiating payments, and performing simple operations, but these capabilities are fragmented. The issue is not whether there are capabilities, but rather that these capabilities have not been integrated and run in a secure, controllable execution environment. So we started to think: If the Agent is a user, should we build an operating system for the Agent? That's our starting point.


Onchain OS: An Operating System Designed for Agents


We don't just consider users as human users; we believe that the Agent itself is also a user. Therefore, we need a system that can serve different types of Agents, allowing them to invoke capabilities, perform operations, and collaborate with each other.


The Onchain OS is not something built from scratch but is built on top of the existing capabilities of the OKX Wallet. The OKX Wallet system itself is API-first; we handle over 1.2 billion daily API calls, with around $3 billion in transaction volume, and currently support over 60 chains, aggregating over 500 DEXs, with an overall system stability of 99.9%. It is because everything is API-ready that we can continue to build on this foundation with the MCP, allowing future Agents to directly invoke these capabilities to accomplish what they want to do.



In this system, we have also introduced payment capabilities, allowing Agents to complete payments and transactions; using X Layer as the execution layer, with extremely low costs, even free; and we have introduced a Plugin mechanism, allowing the community and developers to build more capabilities. We hope that users can perform actions directly in natural language, rather than clicking through layers of interfaces.


For example, if you say, "When ETH drops to $2000, convert it to USDC for me," the Agent will continuously track the market price, use OKX Wallet data to make decisions, and then automatically perform price comparison, liquidity adjustment, validation, and transaction execution when the conditions are met. Therefore, future cryptocurrency users may no longer pay attention to specific interface experiences as they do today because all operations will be distilled into a common language. The underlying security, liquidity, asset management, and various complex processes will all be handled by the native underlying system of OKX Wallet.


But what's even more interesting is that this will not be a world with only one Agent. In the future, there will be many Agents. Your Agent will interact with others' Agents, collaborate with third-party Agents, and even engage in payment and settlement with each other. We will enter an environment of Agent-to-Agent and Agent-to-human interactions. In such an environment, we need an infrastructure to support these payment scenarios, and this is the role of the X Layer. It is a Layer 2 based on the Enhanced OP Stack, supporting fast, low-cost settlement, as well as supporting payments and interactions between Agents.


We believe that in the future, there will be an increasing number of Agents participating in on-chain transactions and interactions, potentially forming an ecosystem on-chain that far exceeds human dominance. Therefore, to achieve this, a truly efficient settlement layer is needed. This is why we want to build an open ecosystem — this is just a start, and we need ecosystem partners to participate, such as continuously expanding different types of Skills.


At the same time, we are continuously opening up the ecosystem, allowing the community to submit Skills, which will be reviewed by AI Agents and the security team before being listed for users to install with one click. We also provide tools such as SDKs and CLIs to help developers build Agents, making them closer to application forms. In Onchain OS and the Plugin Store, we hope to see more real-world use cases, ultimately achieving a scenario where one person can manage a project or even an entire investment portfolio.


Conclusion


Over the years, financial infrastructure has heavily relied on top professionals, requiring the formation of a complete team, including analysts, traders, risk managers, developers, etc., and establishing a standardized investment process, while also complying with anti-money laundering (AML) and other compliance requirements, and needing to operate in synergy with such a system. However, the future may be different. With one person, along with multiple types of Agents, there is an opportunity to perform similar work, perhaps even more efficiently.


As we interact with the on-chain ecosystem, we will no longer need to rely entirely on a team to build everything. Every user may have access to capabilities and competitiveness close to large institutions. This means that ordinary users can possess advantages that only professional institutions used to have. I believe this is a very critical development. This is not only a change happening in our ecosystem, but also explains why AI is playing an increasingly important role in this industry.


We can even imagine that in the future, Agent-driven transactions and interactions could surpass human-dominated markets.


Thank you all. If you are interested in Onchain OS and would like to collaborate, please feel free to contact us.


This article is contributed content and does not represent the views of BlockBeats


Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

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

举报 Correction/Report
Choose Library
Add Library
Cancel
Finish
Add Library
Visible to myself only
Public
Save
Correction/Report
Submit