NVIDIA and Google Cloud Expand AI Developer Support to 100,000 Programmers Through Software Optimizations Hardware Infrastructure and Responsible AI Standards
NVIDIA and Google Cloud have expanded their developer community for AI work to support more than 100,000 programmers. According to an update on their official joint platform, both companies are adding new training resources, software integrations, and infrastructure services to accelerate the creation of AI applications, focusing on optimizing software frameworks and building robust standards for responsible and transparent AI in both cloud and edge deployments.
New training resources heavily utilize open source frameworks such as JAX, which has a new path for users who want to learn how to train and run workloads using JAX across the NVIDIA powered AI Hypercomputer from Google Cloud utilizing MaxText. Another codelab adds NVIDIA Dynamo on Google Kubernetes Engine, helping developers optimize large scale inference with complicated models such as mixtures of experts (MoE) using hardware acceleration.
Hardware improvements for the new developer toolset support multi agent applications. The NVIDIA cuDF library within Google Colab Enterprise provides accelerated data science pipeline work for the developer community, while developers may train agent workflows by leveraging Google DeepMind’s Gemma 4 models with NVIDIA Nemotron's open source models, which are supported across Google Cloud G4 virtual machines that use NVIDIA RTX PRO 6000 Blackwell GPUs using spot instances for low cost of compute or standard runs.
The partnership is making content verification a priority in light of the increasing prevalence of autonomous agents and systems, enabling NVIDIA's direct integration of Google DeepMind's SynthID. The new watermark tool adds digital marks on both the visual content from the NVIDIA Cosmos world foundation models for robot and physical AI training, and images and video for greater transparency while enabling companies to operate agentic systems more responsibly.
This collaborative infrastructure is expected to scale development efforts, from simple prototypes to large enterprise applications. Over the course of the planned technology rollout, the partners plan to add new hardware including instances using the NVIDIA Vera Rubin A5X architecture and Google DeepMind Gemini models. Existing users of this framework include prominent technology companies such as OpenAI, Salesforce, and Crowdstrike, which currently operate AI agents.
