TensorCircuit-NG: A Universal, Composable, and Scalable Platform for Quantum Computing and Quantum Simulation

TensorCircuit-NG is a next-generation quantum software platform connecting quantum physics, artificial intelligence, and high-performance computing through a tensor-native programming paradigm. It unifies quantum circuits, tensor networks, and neural networks in differentiable computational graphs built on machine-learning backends such as JAX, TensorFlow, and PyTorch. The framework supports approximate circuit simulation, analog dynamics, fermion Gaussian states, qudit systems, and scalable noise modeling, while also implementing distributed computing strategies for deep quantum circuits. Its benchmark applications include distributed variational quantum algorithms, quantum machine learning, classical-shadow pipelines, and differentiable tensor-network optimization for many-body physics.