ASUS released enterprise-level Agentic AI solutions and full-stack AI computers on NVIDIA GTC Berlin 2026.
Integrate infrastructure, software, and services for secure and scalable agentic AI.
Berlin, October 21st / PRNewswire / -- ASUS today launched its Enterprise Agentic AI Solution, an integrated full-stack solution that combines AI infrastructure, software, and services to help enterprises from development to production deployment, governance, and expansion of agentic AI. ASUS also showcased its complete agentic AI product portfolio at GTC Berlin 2026, covering data center infrastructure and accelerated computing, AI development and workstations, intelligent edge systems, as well as agent computer, providing a unified foundation for the evolution from data centers to desktops and edges.
Rack-level infrastructure for agentic and AI
As agentic and AI transition from experimentation to production, enterprises need infrastructure that can support continuous perception, reasoning, and multi-step execution in scaled scenarios. ASUS Enterprise Agentic AI Solution has integrated three core capabilities: ASUS AI HUB for proxy registration, governance, and monitoring of entire fleets; NVIDIA AI Enterprise and ASUS AI Software Stack for seamless orchestration and workflow integration; as well as ASUS AI servers and infrastructure, including the full-stack ASUS AI factory platform driven by NVIDIA Vera Rubin NVL72. ASUS claims that this platform, through joint design, can provide 30 times higher throughput and 45 times lower token costs per megawatt for agentic and AI.
These layers collectively provide a secure, reliable, and scalable foundation, while also simplifying infrastructure planning and enterprise integration. ASUS states that through a simplified configuration process based on model scale, performance objectives, and operational requirements, the company can help customers in finance, healthcare, manufacturing, the public sector, and other regulated industries identify and deploy the platforms that best meet their needs.
Enterprise-oriented agentic AI infrastructure
At the corporate level, ASUS showcased a series of high-performance, scalable servers designed for agentic and AI workloads, including XA P2N to E2, which is a 2U server equipped with NVIDIA Vera CPUs and is aimed at next-generation agentic and AI systems; as well as XA P8A to E14AXL, a high-density 6U system that supports up to eight liquid-cooled NVIDIA RTX PRO 6000 Blackwell Server Edition GPU. The ASUS ESC8000 series has expanded its support for high-demand AI and accelerated computing workloads: ESC8000-E12P supports NVIDIA RTX PRO 4500 Blackwell Server Edition GPU for AI, data processing, and visual computing; ESC8000A-E13X combines NVIDIA RTX PRO 6000 Blackwell GPU with NVIDIA ConnectX to SuperNICs to achieve high-speed interconnection.
Extending agentic AI from the data center to the physical world
ASUS demonstrated a practical physical AI workflow that covers simulation, verification, and real-world deployment at the edge. The PE3000N is equipped with NVIDIA Jetson T5000 modules built on NVIDIA Jetson Thor, offering up to 2,070 FP4 TFLOPS of AI performance and 128GB of memory in a compact, deployable form; the PE6000G is a workstation-level AI designed for industrial environments. Together, they support advanced perception, reasoning, and action in applications such as robotics, automated machinery, and intelligent video analysis, and connect development and verification with reliable edge execution within the NVIDIA robotics ecosystem.
Agent computer Expands AI from development to the edge
In addition to data centers, ASUS also provides AI infrastructure driven by NVIDIA at the edge, workstations, and PC endpoints, offering a flexible foundation for enterprises to build, deploy, and scale agentic AI that can be used in production.
ASUS ExpertCenter Pro ET900N G3 brings data center-level performance to the desktop, based on NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip, equipped with 748GB of consistent memory and a complete NVIDIA AI software stack. ASUS claims that this product allows enterprises to securely develop, train, and deploy large language models and generative AI locally, while maintaining full control over the data.
To make the desktop version more user-friendly, agentic AI, ASUS Ascent GX10 is based on the NVIDIA DGX Spark platform and comes equipped with a complete NVIDIA AI software stack. This product currently offers two unified memory configurations: 64GB and 128GB, providing greater flexibility for AI developers across various local AI workloads. This desktop AI supercomputer can support models with approximately 30 to 35 billion parameters, as well as models with up to 200 billion parameters, which can be used for advanced AI development, prototyping, fine-tuning, and inference. As workloads grow, it is possible to easily cluster up to four systems through NVIDIA Sync Cluster Assistant.
Laptops ProArt P16 and P14 equipped with NVIDIA RTX Spark can provide up to petaflop of AI performance, and are fitted with a maximum of 128GB of unified memory. They combine high-performance AI computing with ASUS Lumina Pro OLED displays, with a peak brightness of up to 1,600 nits. The ProArt P16 has a thickness of 12.9 millimeters and a weight of 1.77 kilograms, making it the thinnest 16-inch ProArt laptop to date; the ProArt P14 has a thickness of 13.9 millimeters and a weight of 1.48 kilograms, making it the lightest ProArt laptop currently available. High-capacity batteries – P16 with a maximum of 99.9Wh and P14 with a maximum of 90Wh – support longer periods of mobile creative work. ASUS has also collaborated with Hermes Agent and ComfyUI to bring industry-verified workflows and creative tools that can connect to proxies directly to users' PC, enabling generative AI creation and AI assisted development to run locally prioritously, with flexible access to cloud services when needed.
In order to achieve a constantly online AI on the desktop, ProArt GR1X, mini, and PC integrate NVIDIA RTX performance into a compact 150x150x51 millimeter chassis. This product is equipped with a 218-blade dual-fan cooling system and seven-level fan control, designed to support 24/7 continuous operation with low noise. It supports dual M.2, SSD storage, 10GbE, WiFi, Bluetooth 5.4, and up to four 4K monitors, enabling autonomous agents to process complex 3D rendering and creative workflows in the background overnight.
Welcome to experience ASUS at NVIDIA GTC Berlin 2026.
Asus invites visitors to visit booth number 3039 at Station Berlin during GTC Berlin 2026 to learn about its latest AI solutions in infrastructure, edge computing, and professional applications. For a deeper understanding of the ASUS Edge - to - Cloud AI product portfolio – including edge AI systems as well as enterprise infrastructure based on NVIDIA HGX and RTX PRO – visitors can also go to W3 Hub on ASUS VIP Room. The exhibition is open from October 20th to 22nd. For more information, please visit here.
Supply Situation and Pricing
ASUS server products are available for purchase worldwide. For more details, please contact your local ASUS representative.
