TradeFM: A Generative Foundation Model for Trade-flow and Market Microstructure

2026-02-27

AI Summary

This work proposes a generative Transformer-based approach to modeling market microstructure, designed to learn cross-asset, generalizable representations from massive heterogeneous trading events and synthesize order flows that replicate the statistical properties of real markets. By introducing scale-invariant features and a universal event tokenization scheme, the method maps multi-asset order flows into unified discrete sequences and leverages a deterministic market simulator to generate realistic trading trajectories. Notably, it achieves zero-shot cross-market generalization without asset-specific calibration—the first such foundation model in microstructure modeling. The synthetic data accurately reproduces stylized facts such as heavy-tailed returns and volatility clustering, reducing distributional errors by a factor of 2–3 compared to Compound Hawkes baselines, and demonstrates effective transfer to unseen Asia-Pacific markets.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Machine Learning: Deep Generative Models & Autoencoders
• Computer Vision: Diffusion Models for Vision

Application Category

• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Web data generation and simulation

Abstract

Foundation models have transformed domains from language to genomics by learning general-purpose representations from large-scale, heterogeneous data. We introduce TradeFM, a 524M-parameter generative Transformer that brings this paradigm to market microstructure, learning directly from billions of trade events across>9K equities. To enable cross-asset generalization, we develop scale-invariant features and a universal tokenization scheme that map the heterogeneous, multi-modal event stream of order flow into a unified discrete sequence -- eliminating asset-specific calibration. Integrated with a deterministic market simulator, TradeFM-generated rollouts reproduce key stylized facts of financial returns, including heavy tails, volatility clustering, and absence of return autocorrelation. Quantitatively, TradeFM achieves 2-3x lower distributional error than Compound Hawkes baselines and generalizes zero-shot to geographically out-of-distribution APAC markets with moderate perplexity degradation. Together, these results suggest that scale-invariant trade representations capture transferable structure in market microstructure, opening a path toward synthetic data generation, stress testing, and learning-based trading agents.

Problem

• market microstructure
• trade-flow modeling
• foundation model
• cross-asset generalization
• generative modeling

Innovation

• Foundation Model
• Market Microstructure
• Scale-Invariant Representation
• Generative Transformer
• Order Flow Tokenization