Introduction#
NVIDIA and Google Cloud have announced an expansion of their partnership aimed at enhancing artificial intelligence (AI) capabilities. This announcement was made during the Google Cloud Next event in Las Vegas.
New Infrastructure Details#
The companies introduced the NVIDIA Vera Rubin-powered A5X bare-metal instances, which can support up to 960,000 NVIDIA Rubin GPUs across multiple sites. Within a single site, these A5X instances can utilize up to 80,000 NVIDIA Rubin GPUs. This infrastructure is designed to significantly reduce costs and improve efficiency, offering up to ten times lower costs per token and ten times higher throughput per megawatt compared to previous models.
Google Cloud's NVIDIA Portfolio#
Google Cloud's NVIDIA Blackwell portfolio includes various virtual machines (VMs) such as A4 VMs with NVIDIA HGX B200 systems and A4X VMs with NVIDIA GB200 NVL72 systems. These systems are designed to handle large-scale AI workloads efficiently. Notably, OpenAI is leveraging NVIDIA’s technology on Google Cloud for some of its applications, including ChatGPT.
Advancements in AI Applications#
The collaboration also features new models like Google Gemini, which are now in preview on Google Distributed Cloud. Additionally, the introduction of Confidential G4 VMs with NVIDIA RTX PRO 6000 Blackwell GPUs marks a significant step in confidential computing within the cloud. Companies like CrowdStrike are using NVIDIA’s NeMo libraries for cybersecurity applications, showcasing the practical benefits of this partnership.
Recognition and Future Prospects#
NVIDIA has been recognized as Google Cloud Partner of the Year in two categories: AI Global Technology Partner and Infra Modernization Compute. This acknowledgment highlights the ongoing commitment of both companies to advance AI technology and infrastructure.