Memory-Enforced GPU Sharing in Kubernetes with NVIDIA CUDA 13.4
<img width="768" height="432" src="https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-768x432.webp" class="webfeedsFeaturedVisual wp-post-image" alt="" style="display: block; margin-bottom: 5px; clear:both;max-width: 100%;" link_thumbnail="" decoding="async" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-768x432.webp 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-179x101.webp 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-300x169.webp 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-625x352.webp 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-1536x864.webp 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-645x363.webp 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-660x370.webp 660w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-500x281.webp 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-160x90.webp 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-362x204.webp 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-196x110.webp 196w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-1024x576.webp 1024w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-960x540.webp 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV.webp 1920w" sizes="(max-width: 768px) 100vw, 768px" title="CUDAMemoryKV" />Many AI workloads need only a fraction of a GPU. Giving each one an entire device can leave capacity idle; sharing GPUs lets platform teams serve more workloads...<img width="768" height="432" src="https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-768x432.webp" class="webfeedsFeaturedVisual wp-post-image" alt="" style="display: block; margin-bottom: 5px; clear:both;max-width: 100%;" link_thumbnail="" decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-768x432.webp 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-179x101.webp 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-300x169.webp 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-625x352.webp 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-1536x864.webp 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-645x363.webp 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-660x370.webp 660w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-500x281.webp 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-160x90.webp 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-362x204.webp 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-196x110.webp 196w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-1024x576.webp 1024w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV-960x540.webp 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2026/10/CUDAMemoryKV.webp 1920w" sizes="auto, (max-width: 768px) 100vw, 768px" title="CUDAMemoryKV" /><p>Many AI workloads need only a fraction of a GPU. Giving each one an entire device can leave capacity idle; sharing GPUs lets platform teams serve more workloads with the infrastructure they already have. NVIDIA GPUs already support several approaches to sharing. Multi-instance GPU (MIG) splits a GPU into hardware-isolated instances that are transparent to the application…</p>
<p><a href="https://developer.nvidia.com/blog/memory-enforced-gpu-sharing-in-kubernetes-with-nvidia-cuda-13-4/" rel="nofollow" data-wpel-link="internal" target="_self">Source</a></p>
Read original article ↗
Related Articles
EdgeStart is a self-contained, 8-GPU NVIDIA cabinet that enables building owners to earn money from their available sola
NVIDIA RTX Spark Powers Neural Rendering and AI Characters in Games
NVIDIA RTX Spark is a new Windows PC platform for AI development, creation, and gaming in slim laptops and compact deskt
Your Robot’s Safety Functions Already Work. What If the Input Lies?
This article is brought to you by VicOne. A mobile robot slows for someone in a hallway. A collaborative arm eases its m
Fine-Tuning NVIDIA Nemotron to Know Local Rules, Not Just Local Languages
As AI adoption expands globally, people across languages and cultures are using it in more aspects of their daily lives.