<p>This paper presents HIDE, a novel and high-integrity data embedding method designed for H.266/VVC video streams. HIDE leverages a dynamic carrier control mechanism to achieve efficient, secure, and high-fidelity data hiding while maintaining minimal impact on video quality. The proposed method introduces an entropy-based complexity analysis of Coding Tree Units (CTUs), enabling adaptive classification into high- and low-complexity regions. This classification facilitates intelligent carrier selection and dynamic adjustment of the embedding capacity based on texture characteristics. A Carrier Selection Criterion (CSC) is defined to probabilistically select CTUs for embedding using a reproducible random function initialized with a shared seed. Within each selected CTU, the smallest N PUs with directional intra prediction modes are identified as carriers, forming a structured carrier vector used for data embedding. A key feature of HIDE is its modular architecture, which inherently supports parallel processing and GPU acceleration—making it suitable for deployment in large-scale and real-time video streaming systems where high-performance computing capabilities are essential. A binary stream derived from the secret data is mapped onto this vector through a controlled mode-based modification process, where only one prediction mode is adjusted by ± 1 to embed the data. This minimal change ensures high integrity and imperceptibility, preserving both visual quality and compression efficiency. The data extraction process in H.266/VVC involves extracting hidden data embedded within the compressed video stream, ensuring efficient retrieval of the embedded hidden data. Experimental results demonstrate that HIDE achieves excellent performance in terms of embedding capacity, security, and robustness, making it a promising solution for secure and reliable data embedding applications in modern video coding environments.</p>