Multi-Instance GPU Explained
Multi-Instance GPU matters in hardware work because it changes how teams evaluate quality, risk, and operating discipline once an AI system leaves the whiteboard and starts handling real traffic. A strong page should therefore explain not only the definition, but also the workflow trade-offs, implementation choices, and practical signals that show whether Multi-Instance GPU is helping or creating new failure modes. Multi-Instance GPU (MIG) is an NVIDIA technology introduced with the A100 that allows a single physical GPU to be partitioned into up to seven independent GPU instances. Each instance has its own compute resources, memory, memory bandwidth, and cache, providing hardware-level isolation that guarantees quality of service for each workload.
MIG addresses the common problem of GPU underutilization in inference scenarios. A single large language model inference request may only use a fraction of an A100 or H100's compute capacity. Without MIG, running multiple models on one GPU requires time-sharing with potential interference. MIG provides true hardware isolation, allowing different teams, models, or customers to share a single GPU without performance interference.
MIG instances can be configured in various sizes: on an A100, options include 1g.5gb (1/7th GPU), 2g.10gb (2/7th), 3g.20gb (3/7th), 4g.20gb (4/7th), and 7g.40gb (full GPU). Each instance appears as a separate GPU to software, with its own device ID and compatibility with CUDA, Docker, and Kubernetes. MIG is particularly valuable in cloud and multi-tenant environments.
Multi-Instance GPU is often easier to understand when you stop treating it as a dictionary entry and start looking at the operational question it answers. Teams normally encounter the term when they are deciding how to improve quality, lower risk, or make an AI workflow easier to manage after launch.
That is also why Multi-Instance GPU gets compared with A100, H100, and NVIDIA. The overlap can be real, but the practical difference usually sits in which part of the system changes once the concept is applied and which trade-off the team is willing to make.
A useful explanation therefore needs to connect Multi-Instance GPU back to deployment choices. When the concept is framed in workflow terms, people can decide whether it belongs in their current system, whether it solves the right problem, and what it would change if they implemented it seriously.
Multi-Instance GPU also tends to show up when teams are debugging disappointing outcomes in production. The concept gives them a way to explain why a system behaves the way it does, which options are still open, and where a smarter intervention would actually move the quality needle instead of creating more complexity.
