KDDI launched a joint initiative with Nvidia, Keysight Technologies, and Samsung to develop a high-fidelity radio access network (RAN) digital twin for AI-driven optimization and autonomous network operations in the 6G era. The project creates a virtual replica of real-world RAN conditions to train, test, and validate AI models without disrupting live commercial networks. Nvidia will provide digital twin infrastructure through its Aerial Omniverse Digital Twin and accelerated computing platforms, Keysight will contribute user equipment emulation technology, and Samsung Research America will provide virtualized RAN (vRAN) technology. KDDI will contribute commercial network data and lead trials, while KDDI Research will define technical requirements and develop propagation prediction technologies. One use case includes a training environment for AI-based network optimization; another is a field-trial platform for new AI-based RAN features. KDDI Research added one of its AI propagation models, using a convolutional neural network (CNN) with a residual network (ResNet) structure, improved estimation error in unknown regions by about 15% compared with a conventional AI model. The companies plan to deliver a prototype by the end of March 2028 and aim to expand the platform by the end of 2030.