Jensen Huang Founder, President and CEO at NVIDIA | Official website
NVIDIA and Google Cloud have joined forces in a new collaboration aimed at accelerating the development of generative AI applications and services for startups worldwide. The collaboration, announced at the Google Cloud Next ‘24 event in Las Vegas, seeks to provide support and resources for startups through programs offered by both companies.
Qualified members of NVIDIA Inception, a program supporting over 18,000 startups globally, will now have access to Google Cloud infrastructure with credits of up to $350,000 for those focusing on AI. Additionally, high-growth software makers from both programs can benefit from onboarding to Google Cloud Marketplace, co-marketing, and product acceleration support.
In a joint effort, NVIDIA and Google have optimized the Gemma family of models, developed by Google DeepMind, on NVIDIA AI platforms. This collaboration aims to reduce costs for customers and accelerate innovation for specific use cases in the AI domain.
The partnership also involves making it easier to deploy the NVIDIA NeMo framework across Google Cloud platforms, allowing developers to automate and scale the training and serving of generative AI models effectively. Furthermore, the availability of NVIDIA-accelerated generative AI computing is set to expand with the upcoming release of A3 Mega instances on Google Cloud, powered by NVIDIA H100 Tensor Core GPUs.
Looking ahead, NVIDIA’s upcoming GPUs based on the NVIDIA Blackwell platform, namely the NVIDIA HGX B200 and the NVIDIA GB200 NVL72, are scheduled to be integrated into Google Cloud early next year. The GB200 NVL72, designed for massive-scale model training and real-time inferencing, promises significant advancements in performance, bandwidth, and communication capabilities.
NVIDIA's DGX Cloud, an AI platform optimized for generative AI demands, will be available on A3 VMs powered by H100 GPUs and the GB200 NVL72 on Google Cloud in 2025, offering enhanced capabilities for enterprise developers.
