	A100	A30	L40	L4	A16
GPU Architecture	NVIDIA Ampere	NVIDIA Ampere	NVIDIA Ada Lovelace	NVIDIA Ada Lovelace	NVIDIA Ampere
Memory Size	80GB / 40GB HBM2	24GB HBM2	48GB GDDR6 with ECC	24GB GDDR6	64GB GDDR6 (16GB per GPU)
Virtualization Workload	Highest performance virtualized compute, including AI, HPC, and data processing. Includes support for up to 7 MIG instances. Upgrade path for V100/V100S Tensor Core GPUs.	Virtualize mainstream compute and AI inference, includes support for up to 4 MIG instances. Upgrade path for T4.	High-end 3D visualization applications, AI training, and inference workloads. Upgrade path for Quadro® RTX 8000, Quadro RTX 6000, or T4.	Entry-level to mid-range design and creative workloads for virtual workstations. Video streaming, office productivity tools for virtual desktops. Upgrade path for T4.	Office productivity applications, streaming video and teleconferencing tools for graphics-rich virtual desktops accessible from anywhere. Upgrade path for M10 or T4.
vGPU Software Support	NVIDIA AI Enterprise	NVIDIA AI Enterprise	NVIDIA RTX Virtual Workstation (vWS), NVIDIA Virtual PC (vPC), NVIDIA Virtual Apps (vApps), NVIDIA AI Enterprise^*	NVIDIA RTX vWS, vPC, vApps, NVIDIA AI Enterprise	NVIDIA RTX vWS, vPC, vApps, NVIDIA AI Enterprise

GPUs for Virtualization

Compare GPUs for Virtualization