A. It combines a CPU and a GPU on a single superchip, connected by a high-speed, low-latency NVLink-C2C interconnect.

B. It is the first NVIDIA GPU to feature fourth-generation Tensor Cores for enhanced matrix calculations.

C. It uses on-chip encryption to create a confidential computing environment for the CPU.

D. It replaces the need for server memory (DRAM) by using the GPU's global memory exclusively.