Google's Gemini AI Now Runs On-Prem

Google Distributed Cloud (GDC) is a fully managed hardware and software solution that extends Google's cloud infrastructure to on-premises data centers and edge locations. This allows organizations to run services like Gemini AI locally, addressing needs for data residency, low-latency processing, and regulatory compliance. The platform is offered in both a "connected" version and a fully "air-gapped" configuration that requires no connection to the public internet. The on-premises solution is a result of a partnership between Google and NVIDIA, leveraging NVIDIA's Blackwell and Hopper GPU architectures to run Gemini models. This collaboration enables enterprises to utilize the advanced reasoning and multimodal capabilities of Gemini, which can process text, images, audio, and video, directly within their own secure environments. The public preview for Gemini on GDC was slated for the third quarter of 2025. For industries like logistics and retail, on-premise AI offers significant advantages in operational efficiency and customer experience. AI-powered tools can optimize inventory management and demand forecasting with greater accuracy by analyzing real-time data, market trends, and even weather patterns. In logistics, this translates to optimized delivery routes, reduced fuel consumption, and more accurate inventory counts, which can save companies hundreds of thousands of dollars annually. The GDC air-gapped option provides a high level of security for sensitive data, having received authorization for use in U.S. Government Secret and Top Secret missions. This configuration ensures that all data processing occurs in complete isolation, meeting stringent sovereignty and compliance standards such as NIST SP 800-53. The platform supports confidential computing on both CPUs and GPUs to encrypt data even while it's being processed. Beyond Gemini, Google Distributed Cloud supports a range of other AI tools and services. This includes Vertex AI for managing machine learning models and other services like Translation API, Speech-to-Text, and optical character recognition (OCR). The platform also supports Google's open-source Gemma models, providing flexibility for various AI-driven applications. Early adopters of Gemini on GDC include government agencies in Singapore and the Japanese telecom provider KDDI Corp. These organizations highlight the benefit of innovating with advanced AI while adhering to local compliance and data residency requirements. The system allows for building applications like automated document summarization, intelligent chatbots, and AI-assisted code generation, all while keeping sensitive information secure.