Achieving Secure Generative AI Utilization in On-Premise Environments to Protect Confidential Information. Scuti Publishes a White Paper Summarizing RAG System Verification Results in Collaboration with HPE
Scuti Co., Ltd. partnered with HPE to verify the effectiveness of generative AI and RAG in an air-gapped on-premise environment, releasing a white paper detailing how enterprises can achieve secure, cloud-level AI performance without data leakage risks.
📋 Article Processing Timeline
- 📰 Published: April 6, 2026 at 19:00
- 🔍 Collected: April 6, 2026 at 10:30
- 🤖 AI Analyzed: April 21, 2026 at 02:16 (351h 45m after Collected)
## Demonstrating the Effectiveness of Generative AI and RAG in an Air-Gapped Network Environment
Scuti Co., Ltd. has conducted verification of a Generative AI (GAI) and Retrieval-Augmented Generation (RAG) system in a completely air-gapped on-premise environment, completely isolated from external networks, utilizing the latest server chassis "HPE ProLiant DL380 Gen11" from Hewlett Packard Enterprise Japan (hereinafter referred to as HPE). Today, we are pleased to announce the publication of a technical white paper summarizing the verification process and its results.
In recent years, while the introduction of generative AI is progressing in many companies, there are not a few cases where the use of cloud-based AI must be restricted due to the risk of leaking confidential or personal information. In this verification, we clarified how to achieve high-precision AI utilization in a completely offline environment to address the challenges of management, human resources, and DX promotion departments that require high-level security.
Download the white paper
## Background of the Verification: Eliminating the Trade-off Between Security and Convenience
To truly utilize generative AI for operational efficiency, it is essential to build a "RAG (Retrieval-Augmented Generation)" system that allows AI to learn and refer to the company's unique knowledge (manuals, regulations, past meeting minutes, etc.). However, this information is the most important secret for a company, and there are high psychological and regulatory hurdles to uploading it to a cloud environment.
To solve this problem, our company leveraged HPE's powerful hardware foundation to verify an "On-Premise Secure GAI" that performs absolutely no external communication.
## Adopting the Robust Infrastructure Foundation "HPE ProLiant DL380 Gen11"
For this verification equipment, we adopted the "HPE ProLiant DL380 Gen11," which features outstanding processing capability and reliability.
- High Computing Power: Confirmed performance capable of withstanding vector searches in RAG systems and local execution of Large Language Models (LLMs).
- Robust Security: By using HPE products equipped with hardware-level security features such as Silicon Root of Trust, we ensured consistent safety from the infrastructure layer to the application layer.
Through this verification, we have demonstrated that even in an environment isolated from the outside, it is possible to achieve response speeds comparable to cloud-based AI and generate accurate answers based on the organization's specific expertise.
Scuti Co., Ltd. has conducted verification of a Generative AI (GAI) and Retrieval-Augmented Generation (RAG) system in a completely air-gapped on-premise environment, completely isolated from external networks, utilizing the latest server chassis "HPE ProLiant DL380 Gen11" from Hewlett Packard Enterprise Japan (hereinafter referred to as HPE). Today, we are pleased to announce the publication of a technical white paper summarizing the verification process and its results.
In recent years, while the introduction of generative AI is progressing in many companies, there are not a few cases where the use of cloud-based AI must be restricted due to the risk of leaking confidential or personal information. In this verification, we clarified how to achieve high-precision AI utilization in a completely offline environment to address the challenges of management, human resources, and DX promotion departments that require high-level security.
Download the white paper
## Background of the Verification: Eliminating the Trade-off Between Security and Convenience
To truly utilize generative AI for operational efficiency, it is essential to build a "RAG (Retrieval-Augmented Generation)" system that allows AI to learn and refer to the company's unique knowledge (manuals, regulations, past meeting minutes, etc.). However, this information is the most important secret for a company, and there are high psychological and regulatory hurdles to uploading it to a cloud environment.
To solve this problem, our company leveraged HPE's powerful hardware foundation to verify an "On-Premise Secure GAI" that performs absolutely no external communication.
## Adopting the Robust Infrastructure Foundation "HPE ProLiant DL380 Gen11"
For this verification equipment, we adopted the "HPE ProLiant DL380 Gen11," which features outstanding processing capability and reliability.
- High Computing Power: Confirmed performance capable of withstanding vector searches in RAG systems and local execution of Large Language Models (LLMs).
- Robust Security: By using HPE products equipped with hardware-level security features such as Silicon Root of Trust, we ensured consistent safety from the infrastructure layer to the application layer.
Through this verification, we have demonstrated that even in an environment isolated from the outside, it is possible to achieve response speeds comparable to cloud-based AI and generate accurate answers based on the organization's specific expertise.