PlanX: Reconstructing On-Chain Execution with AI, Moving Towards a New Paradigm

Mar 2, 2026 17:30:08

Share to

1. A New Operational Paradigm for On-Chain Finance

As the centralized financial system continues to evolve into a decentralized financial system, the on-chain market is entering a brand new operational paradigm: high-frequency, complex, operating 24/7, and gradually dominated by automated decision-making systems and AI execution models.

In this paradigm, trading is no longer centered around manual operations or interface interactions, but is continuously run by algorithms, strategy systems, and execution engines. The speed of market state changes, information density, and structural complexity have far exceeded the limits that traditional manual trading and platform interaction models can bear.

2. Structural Bottlenecks of Traditional Trading Models

In the current on-chain environment, the traditional "platform"-centered trading model is showing obvious structural limitations:

Natural limitations of humans in terms of time scale
Cognitive bandwidth cannot cover high-frequency, multi-state parallel market changes
Manual execution has inevitable fluctuations in stability, consistency, and long-term operation

As the market enters a phase dominated by automated systems, the interaction method between humans and platforms is itself becoming a bottleneck for execution efficiency.

3. The Birth Background of PlanX

PlanX was born in this structural transformation.

PlanX is an AI-centric on-chain execution protocol dedicated to upgrading trading execution from "platform behavior" to a protocol-level, verifiable, evolvable, and long-term operational execution infrastructure, thereby reconstructing the collaboration between humans and the on-chain financial system.

Execution Beyond Human is not about replacing humans, but about expanding the boundaries of human capabilities within the financial system.

Founding Team

PlanX was founded by a technical team known for its engineering capabilities, system design experience, and practical on-chain execution. The core consensus of the team is very clear: the essence of competition in on-chain finance is not marketing, traffic, or short-term arbitrage, but the engineering quality and long-term stability of the execution system.

Lex Li | Co-founder & CEO

Lex Li graduated from UCLA with a degree in Electrical Engineering and Integrated Circuits (EEIC), possessing a diverse background spanning aerospace engineering, communication systems, algorithmic trading, and blockchain entrepreneurship.

He has participated in the development of high-performance, mission-critical systems for SpaceX and AT&T, with a deep engineering understanding of the operation of low-latency, high-reliability systems under extreme conditions. After entering the Web3 space, Lex has continued to focus on token incentives and protocol-level economic model design, the underlying structure of decentralized financial systems, on-chain/off-chain collaborative execution architecture, crypto asset strategies, and system-level risk management.

Through multiple system implementations and entrepreneurial practices, Lex has gradually formed a methodology that derives financial structures from engineering constraints, which has also become the core ideological source for PlanX's execution layer design.

Michael Gao | Co-founder & CTO

Michael Gao graduated from UC Berkeley with a degree in Electrical Engineering and Computer Science (EECS) and has worked at DELL EMC, McAfee, Taraxa, and InfStones, with over 10 years of software and systems engineering experience.

He has long focused on blockchain infrastructure and cryptography, high-concurrency asynchronous distributed systems, enterprise-level security and infrastructure engineering, Layer 1 public chains, PoS consensus, and multi-chain staking systems.

Michael has been deeply involved in the core design and implementation of multiple public chains and infrastructure projects, possessing the ability to understand on-chain state machines, security boundaries, execution determinism, and performance bottlenecks at a system level. At PlanX, he is responsible for structuring complex trading, settlement, and risk control logic into a verifiable, scalable, and long-term operational on-chain execution protocol.

Technical Advisors

PlanX also receives support from two long-term AI research and engineering advisors from DeepMind and Waymo. Their professional backgrounds mainly focus on large-scale machine learning systems, automated decision-making and reinforcement learning, high-reliability automated execution frameworks, and human-machine collaboration and autonomous control in complex systems.

These experiences provide important theoretical and engineering support for PlanX in AI-driven execution, agent architecture design, and the evolution of long-term autonomous systems.

Engineering Philosophy

The PlanX team does not position itself as a traditional "trading platform," but as builders of on-chain execution systems.

Execution-First System Design Principles

The team always adheres to the following core principles:

Define execution constraints first, then design financial behaviors
Use systems engineering methods to break down trading, risk, and liquidity issues
Transform uncertain human behaviors into verifiable protocol-level execution logic

From Platform Logic to Execution Infrastructure

In the architecture of PlanX:

Execution is not a UI layer experience issue, but a protocol layer determinism issue
Risk control is not parameter tuning, but state machine and constraint design
AI is not a gimmick, but an automated extension of the execution layer

This allows PlanX to build products from the perspective of Execution Infrastructure from the very beginning, rather than replicating existing DEX or trading platform models.

Xgent

Within the overall system of PlanX, Xgent is not a single strategy model or trading tool, but a vertical intelligence layer aimed at future financial forms.

Core Goal: Countering Institutional-Level AI

The core goal of Xgent is to provide execution capabilities that can counter institutional-level AI for all retail traders and trading platforms.

In the context of the gradual arrival of Web 4.0, the main adversary in trading is no longer "humans," but large-scale AI trading models deployed by institutions. The challenges faced by retail and small platforms have shifted to execution speed, strategy combination capabilities, systematic risk control, and real-time response capabilities. Xgent is designed to address this generational asymmetry.

Natural Language → Strategy Agent Architecture Output

Xgent's long-term vision is to make "natural language input → strategy agent architecture output" a fundamental capability for traders, rather than an exclusive privilege of institutions.

Traders only need to express their goals, constraints, and risk preferences, and Xgent can transform them into executable, verifiable, and long-term operational strategy systems. This is not about replacing human decision-making, but about amplifying human judgment and risk awareness, leaving execution and optimization to machines.

Countering Institutional AI Trading Models, Not Imitating

Xgent rejects opaque, unverifiable black-box models, emphasizing open strategy expression, modular agent architecture, and assessable, evolvable execution intelligence.

When institutions rely on large-scale AI, individuals can still achieve equal or even superior execution capabilities through structured intelligence.