Head-to-head against traditional GPUs like Nvidia’s H100, Kneron’s new Kneo Rack cluster provides same level of task performance while using 60% less energy
San Diego, USA – January 6, 2026 — Kneron, a global leader in full-stack edge AI solutions and NPU chip pioneer, today announced the completion of its KneoEdge™ enterprise AI infrastructure platform with the launch of Kneo Rack, a rack-mounted AI inference cluster purpose-built for data centers and large-scale deployments.
The new Kneo Rack cluster delivers data-center-class performance without the traditional energy burden. Compared to GPU servers such as NVIDIA H100, Kneo Rack reduces power consumption by 60% while achieving the same level of task performance – directly addressing enterprise challenges around energy efficiency, thermal management, and data sovereignty.
Also including Kneo 300 (entry-level edge AI appliance) and Kneo 350 (high-performance edge server) as well as Kneo Rack, the KneoEdge™ portfolio marks Kneron’s strategic entry into the enterprise AI infrastructure market. KneoEdge™ offers truly trusted, usable, and controllable on-premise AI deployment solutions for industries with stringent data security and compliance requirements—including finance, healthcare, manufacturing, energy, and smart cities.
“Our vision is simple: every organization that cares about data sovereignty should have its own AI infrastructure,” said Dr. Albert Liu, CEO of Kneron. “KneoEdge™ isn’t just a product—it’s a new AI paradigm: advanced intelligence doesn’t need the cloud, security is the baseline, and sovereignty is non-negotiable.”
As generative AI transforms enterprise operations, organizations face a critical challenge: how to harness AI’s power without sacrificing data sovereignty, compliance, or cost control. While public cloud AI offers convenience, it falls short for high-trust, regulated industries with stringent security requirements.
KneoEdge™ resolves this “AI deployment paradox” by delivering cutting-edge AI capabilities entirely on-premise—no cloud connectivity required. The platform integrates Kneron’s proprietary NPU chips, a hardened secure operating system (KneoOS), an optimized AI inference engine, and centralized management into a unified architecture where data stays local, intelligence runs at the edge, and management is unified.
Introducing Kneo Rack: The Private AI Cloud, Reimagined
Kneo Rack targets headquarters data centers, large institutions, and enterprises deploying AI at scale. Typical use cases include:
- Real-time event monitoring and personnel behavior analysis in large manufacturing plants
- Automated container identification and scheduling at major ports
- Mission-critical security systems for banks and government agencies
- Enterprise-wide AI inference serving multiple departments and locations
Traditional GPU-based inference servers deliver power but come with crippling operational costs: massive electricity consumption, complex liquid cooling requirements, and strict data center environmental specifications that often necessitate facility retrofits or new construction.
Kneo Rack leverages Kneron’s heterogeneous NPU architecture to deliver comparable AI inference throughput at revolutionary efficiency levels:
- Dramatically lower power consumption: The total system power draw is significantly less than comparable GPU servers (e.g., NVIDIA H100), reducing electricity costs by over 60% and slashing long-term operational expenses.
- No liquid cooling required: Operates reliably with standard air cooling—eliminating cold aisle containment, chiller plants, and precision HVAC systems
- Deploy anywhere: Functions in standard office server rooms or edge facilities without specialized infrastructure
- Lower Total Cost of Ownership (TCO): From power bills and HVAC load to UPS capacity and rack space, Kneo Rack optimizes every aspect of infrastructure. Enterprises avoid building new high-spec data centers or bearing high PUE (Power Usage Effectiveness) penalties—truly achieving “plug in and get intelligent.”
Jointly developed with Quantea, a Silicon Valley infrastructure partner, Kneo Rack integrates multiple Kneo 350 servers under Quantea’s enterprise-grade management controller, paired with an intuitive visual application orchestration platform. The fully integrated system includes cluster management and pre-packaged software, eliminating the need for extensive DevOps resources or custom integration work.
The Complete KneoEdge™ Portfolio
With Kneo Rack now available, KneoEdge™ offers three deployment tiers to match any enterprise requirement:
- Kneo 300: Entry-level edge AI appliance for branch offices, retail locations, and light workload environments. Plug-and-play design with ultra-low power consumption and privacy-first local processing.
- Kneo 350: High-performance edge server for factories, hospitals, and regional operations centers. Delivers 4× the compute performance of Kneo 330 with pre-loaded applications like KneoChat and Kneron’s AI Agent development framework for rapid custom workflow creation.
- Kneo Rack: Rack-scale AI inference cluster for data centers and enterprise-wide deployments. Delivers sustainable, scalable on-premise AI with simplified thermal management and industry-leading energy efficiency.
Beyond compute power, KneoEdge™ delivers six ready-to-use industry applications, allowing customers to immediately benefit from localized AI tools:
- KneoChat™ RAG: Retrieval-Augmented Generation chatbot that securely queries enterprise documents (PDFs, Word, Excel, PPT) stored locally to deliver accurate, context-aware answers grounded in real company data.
- KneoMeet™: AI meeting assistant that automatically records, transcribes, summarizes, and extracts action items from Microsoft Teams, Zoom, and Google Meet—with all audio processed on-premise for compliance.
- KneoSurvey™: Healthcare-specific assistant for secure patient data collection, pre-visit screening, symptom assessment, and clinical data integration—all with end-to-end encryption and local processing.
- KneoAnalyze™: Natural language data analytics engine that lets non-technical users query databases and files (SQL Server, Oracle, MySQL, Excel, CSV) through conversational prompts without IT dependency.
- KneoVision™: AI video understanding engine for real-time surveillance, intrusion detection, safety compliance monitoring (PPE), crowd analytics, and equipment anomaly detection—with all video processing on-device.
- KneoSense™: Multimodal AI assistant that combines speech, text, and image understanding with task automation, IoT integration, and personalized learning—all while keeping behavioral data on-premise.
Kneron plans to expand the KneoEdge™ ecosystem in 2026 with the Kneo LLM series, enabling local inference of large language models. The company will also deepen integrations with leading OS, database, and middleware partners to ensure seamless enterprise adoption.