ZOTAC Showcases AI-Accelerated Embedded and Enterprise Solutions at COMPUTEX 2025

At COMPUTEX 2025, ZOTAC Technology presented a consolidated view of its enterprise and embedded hardware roadmap, emphasizing solutions designed for AI acceleration, edge computing, and large-scale GPU infrastructure. With a focus on sustained product availability, thermal efficiency, and modular scalability, ZOTAC’s lineup addressed current demands in AI model deployment, inference workloads, and embedded industrial systems.

Embedded and Edge AI with ZBOX PRO

ZOTAC showcased its latest ZBOX PRO Embedded Solutions, purpose-built for AI edge computing and industrial applications. Featured products include:

• ZBOX PRO AI BOX Mini PCs – These systems incorporate add-on AI accelerators from DeepX and Axelera, delivering up to 856 AI TOPS while maintaining energy efficiency and compact deployment footprints.
• NVIDIA Jetson-powered units and Rockchip SoC-based AIPCs, optimized for intelligent edge applications including surveillance, smart retail, and autonomous systems.

• ZOTAC ZBOX PRO ZP-ZU27B4000: Embedded AI PC with NVIDIA RTX PRO Blackwell GPU, is a compact, embedded industrial PC from ZOTAC’s ZBOX PRO line, engineered for AI-accelerated workloads and advanced visualization. It features the NVIDIA RTX PRO 4000 Blackwell Generation embedded GPU, delivering workstation-class graphics performance in a space-efficient design.

These embedded systems are built with extended product life cycles, hardened chassis, and features such as Hardware Watchdog, addressing requirements not typically met by consumer-grade mini PCs.

Industrial and Passive-Cooled Solutions

ZOTAC expanded its ZBOX C and M Series with updated passive and ultra-slim designs that target industrial computing, edge inferencing, and low-power AI workloads.

• The ZBOX CI675 nano and CI655 nano feature Intel Core Ultra Series 2 Processors with integrated Neural Processing Units (NPUs), designed for sustained, efficient on-device AI execution.

• The ZBOX edge CI345, featuring the Intel Twin Lake N150 processor, leverages passive cooling for silent, fanless deployments in power-sensitive environments.

• The ZBOX edge MI676 and MI656 introduce active-cooled ultra-slim platforms with Intel Arc Graphics and NPUs, targeting AI-enhanced workloads where size and thermal profile are key constraints.

These systems are suited for industrial automation, medical equipment, retail analytics, and remote sensing deployments—use cases where reliability, thermals, and long-term availability are critical.

Enterprise-Grade GPU Infrastructure

In addition to embedded systems, ZOTAC displayed rack-mounted GPU servers designed for AI training, LLM fine-tuning, and inference clustering. These include:

• 4U and 6U form factor servers, supporting configurations of up to 10 GPUs per chassis, optimized for compute density and thermal performance
• Modular customization options for customers requiring fine-grained control over performance-per-watt, airflow, and memory-to-GPU ratios

These systems are intended for organizations building out internal AI infrastructure—such as research labs, data centers, CSPs, and on-premise enterprise deployments—where flexibility and hardware control are prerequisites.

Embedded MXM Modules with NVIDIA Blackwell

ZOTAC also introduced a new series of embedded MXM GPU modules based on the NVIDIA Blackwell architecture. These GPUs are designed for integration into OEM hardware and industrial platforms requiring AI inference acceleration in constrained form factors.

The new Blackwell-based MXM lineup will support extended lifecycle availability, software compatibility with NVIDIA’s AI stack, and form factor durability suited for field-deployed or ruggedized environments.

Conclusion

ZOTAC’s COMPUTEX 2025 enterprise and embedded showcase demonstrated the company’s continued investment in modular AI infrastructure, passively cooled industrial systems, and long-lifecycle embedded hardware. These offerings reflect the increasing demand for scalable, decentralized AI systems—from edge inferencing to centralized LLM training—and position ZOTAC as a relevant player beyond the consumer GPU segment.

For organizations seeking performance with configurability and long-term hardware support, ZOTAC’s embedded and enterprise solutions represent a practical option for building AI-ready infrastructure across sectors.