Due to the growing number of advanced attacks and escalating frequency of cyber threats, an intelligent, decentralized and tamper-resistant CTI (Cyber Threat Intelligence) mechanism is in more demand than ever before. This article presents an AI framework with blockchain to integrate Federated Learning (FL) and Generative Adversarial Network (GAN) for secure, privacy preserving, adaptive threat intelligence sharing. The proposed system utilizes blockchain’s immutability and transparency to realize the trusted CTI sharing, while AI-empowered analytics drive automated threat validation, anomaly detection and predictive defense. Its smart contract technological foundation and real-time intelligence exchange can enable distributed stakeholders from the government, industry, and research to engage in autonomous response coordination. Experiments with real-world cyber-attacks datasets show the robustness of our method against adversarial perturbations, less false positives, and an increase in detection accuracy when compared to state-of-the-art centralized CTI platforms. Both Advance Persistent Threats (APTs) and zero-days are significantly reduced via the shared, verifiable threat learning employed by the system scalability, compliance with regulations and interoperability of deploying the blockchain-AI CTI systems in large-scale cybersecurity infrastructures is discussed.