Intel and Nutanix Unveil Turnkey Enterprise RAG Platform
Intel and Nutanix unveiled a new turnkey platform for enterprise Retrieval-Augmented Generation (RAG). The solution is designed to allow businesses to securely build and operate Large Language Models using their own proprietary data. The platform aims to simplify the deployment of generative AI within a secure, private environment.
The solution integrates Nutanix's GPT-in-a-Box 2.0 software with Intel hardware, including Xeon processors with built-in AI acceleration and potentially Intel's Gaudi AI accelerators. This avoids the cost and complexity of GPU-dependent infrastructures for many RAG and inference workloads. The platform is designed to provide a full-stack, enterprise-grade environment for securely deploying AI applications on-premises, at the edge, or in the cloud. Retrieval-Augmented Generation is critical for enterprise AI as it grounds large language models on an organization's own, real-time data, reducing hallucinations and improving factual accuracy. This allows the AI to provide responses based on internal knowledge bases, such as technical manuals or policy documents, rather than relying solely on its original, generalized training data. The approach enhances security by keeping proprietary data within the enterprise's control, a key concern for government and defense contractors. The Nutanix Enterprise AI platform, a core component, simplifies the lifecycle management of AI models and applications with a user-friendly interface. It allows IT administrators to manage AI deployments, including monitoring performance and token usage, without needing specialized AI expertise. The platform supports validated LLMs from partners like NVIDIA and Hugging Face, or customers can upload their own models. This partnership reflects a broader industry trend of creating open AI ecosystems, avoiding vendor lock-in associated with vertically integrated stacks. Just days ago, Nutanix announced a similar strategic partnership with AMD to develop an open, full-stack AI platform using AMD EPYC CPUs and Instinct GPUs. This focus on open standards and multiple hardware options provides more flexibility for enterprise customers. For defense contractors, on-premise RAG solutions address significant data sovereignty and security requirements. By processing data locally, sensitive information is not exposed to public AI services, mitigating risks of data leakage and enabling compliance with regulations. The ability to use existing, CPU-based infrastructure can also lower the barrier to entry for deploying AI, avoiding large, upfront investments in specialized GPU hardware.
