Google and NVIDIA have expanded their strategic partnership to drive the future of artificial intelligence, focusing on high-performance infrastructure, cutting-edge AI models, and developer empowerment.
A key highlight is the deep integration of NVIDIA’s Blackwell GPU architecture into Google Cloud. Google is the first cloud provider to offer the powerful NVIDIA HGX B200 and GB200 NVL72 GPUs via its new A4 and A4X virtual machines.
These are available through services like Vertex AI and GKE, providing massive scale and performance for AI workloads, including Google’s Gemini models.
The partnership also covers joint development of community-driven software like JAX, OpenXLA, MaxText, and llm-d, which support open and proprietary AI models.
Gemini, Google’s most advanced AI family, and the lightweight Gemma models have been optimized to run efficiently on NVIDIA GPUs using TensorRT-LLM and are deployable via NVIDIA NIM microservices.
For customers with strict data governance needs such as in finance, healthcare, and government, Gemini models can now be deployed securely on-premises using Google Distributed Cloud and NVIDIA Blackwell. This ensures privacy while unlocking the power of agentic AI.
Beyond infrastructure, Google and NVIDIA are investing in developer ecosystems, including the launch of a joint community and tools that help AI workloads scale across massive GPU clusters using open frameworks.
Together, the companies aim to democratize AI development, enabling faster innovation across industries with reliable performance and secure deployment options.
Shahriena Shukri is a journalist covering business and economic news in Malaysia, providing insights on market trends, corporate developments, and financial policies. More about Shahriena Shukri.
