JFrog announced at Computex 2025 that its software supply chain platform has been integrated with the NVIDIA Enterprise AI Factory-certified architecture. As part of this collaboration, JFrog has positioned itself as the main software component for next-generation agentic AI applications.
Based on the NVIDIA Blackwell platform and targeting in-house AI infrastructures, the system offers a full-stack MLOps solution covering the development, secure management and distribution processes of AI-supported applications.
Fully integrated software chain platform ready for Agentic AI
JFrog’s integrated platform acts as a single source for software components within NVIDIA Enterprise AI Factory. This structure scans organizations’ AI models, software engines and other components for security vulnerabilities, enabling versioning, auditing and monitoring throughout the entire lifecycle.
In addition, all AI models, containers and dependency files developed can be pulled, loaded and distributed directly from the JFrog Platform. This system eliminates the need to pull components from external sources at runtime, providing a closed-loop, secure structure.
In the announcement, it was shared that the JFrog Platform, which runs natively on NVIDIA Blackwell systems, provides low latency, high efficiency, and large-scale processing power. The platform supports a wide range of use cases from autonomous decision-making systems to physical artificial intelligence solutions in many sectors including finance, healthcare, telecom, retail, media, and manufacturing.
Supported by NVIDIA’s engineering power and global partner ecosystem, this architecture helps businesses deploy artificial intelligence faster and reduce operational risk.
JFrog CEO Shlomi Ben Haim said that the company has integrated security, governance, and lifecycle principles, which have long been applied in the software development world, into AI models in order to scale artificial intelligence.
Emphasizing that ML models should be considered as first-class components in the software world, Ben Haim stated that JFrog has assumed the position of being a “single source of truth” for both software and artificial intelligence components with this understanding. With the integration, version tracking of ML models, transition processes, and rapid deployment of updated versions can also be managed from a single center.
NVIDIA’s Vice President of Enterprise AI Software Products, Justin Boitano, said that the system offers institutions a structure based on governance and trust, as well as high performance, and that this facilitates scalable AI innovation.
It was also shared that the integration expanded previous work with NVIDIA’s RTX NIM software infrastructure. In addition, the RTX Remix platform, which uses technologies such as DLSS and ray tracing through the system developed by JFrog, also offers updated image quality examples for old games.
The new system provides local data control and a secure working environment for physical AI systems and high-performance computing (HPC) workloads, especially agentic AI. JFrog’s platform offers a holistic solution covering the entire process from the development to the operation of software components and AI models. As part of the integration, both JFrog’s own test infrastructure and NVIDIA’s verification systems are actively used.
