NVIDIA Introduces DGX Cloud Serverless Inference for Scalable AI Solutions

Ted Hisokawa
Mar 19, 2025 06:22

NVIDIA unveils DGX Cloud Serverless Inference, a new AI solution enabling seamless deployment across cloud environments with enhanced scalability and flexibility, targeting Independent Software Vendors (ISVs).

NVIDIA has announced the launch of DGX Cloud Serverless Inference, a groundbreaking auto-scaling AI inference solution designed to streamline application deployment across diverse cloud environments. This innovative platform aims to simplify the complexities faced by Independent Software Vendors (ISVs) when deploying AI applications globally, according to NVIDIA’s official blog.

Revolutionizing AI Deployment

Powered by NVIDIA Cloud Functions (NVCF), DGX Cloud Serverless Inference abstracts multi-cluster infrastructure setups, allowing for seamless scalability across multi-cloud and on-premises environments. The platform provides a unified approach to deploying AI workloads, high-performance computing (HPC), and containerized applications, enabling ISVs to expand their reach without the burden of managing complex infrastructures.

Benefits for Independent Software Vendors

The serverless inference solution offers several key benefits for ISVs:

• Reduced Operational Complexity: ISVs can deploy applications closer to customer infrastructures with a single, unified service, regardless of the cloud provider.
• Increased Agility: The platform allows for rapid scaling to accommodate burst or short-term workloads.
• Flexible Integration: Existing compute setups can be integrated using bring your own (BYO) compute capabilities.
• Exploratory Freedom: ISVs can trial new geographies and providers without committing to long-term investments, supporting diverse use cases like data sovereignty and low latency requirements.

Supporting Diverse Workloads

DGX Cloud Serverless Inference is equipped to handle a variety of workloads, including AI, graphical, and job workloads. It excels in running large language models (LLMs), object detection, and image generation tasks. The platform is also optimized for graphical workloads such as digital twins and simulations, leveraging NVIDIA’s expertise in graphical computing.

How It Works

ISVs can begin using DGX Cloud Serverless Inference by utilizing NVIDIA NIM microservices and Blueprints. The platform supports custom containers, allowing for autoscaling and global load balancing across multiple compute targets. This setup enables ISVs to deploy applications efficiently, leveraging a single API endpoint for managing requests.

Pioneering Use Cases

Several ISVs have already adopted DGX Cloud Serverless Inference, showcasing its potential to transform various industries. Companies like Aible and Bria are leveraging the platform to enhance their AI-powered solutions, demonstrating significant improvements in cost efficiency and scalability.

As NVIDIA continues to innovate in AI and cloud computing, DGX Cloud Serverless Inference represents a significant step forward in enabling ISVs to harness the full potential of AI technologies with ease and efficiency.

Image source: Shutterstock