ASUS has launched the Ascent QN10, a compact desktop built around Qualcomm's Snapdragon X2 Elite platform with an integrated 80 TOPS Hexagon NPU. The system pairs an 18-core Qualcomm Oryon CPU with integrated Adreno graphics and targets creators, developers, and enterprises running AI-assisted workflows.

The chassis measures under 0.7 liters. According to ASUS, that makes it 86% smaller than conventional 5-liter desktop systems. The company positions the device for home offices, creative studios, and enterprise deployments where desk space and power draw are constraints.

The integrated Qualcomm Adreno GPU delivers up to 2.3 times the performance of the previous-generation graphics solution, ASUS says.

On-device AI processing

The Hexagon NPU provides up to 80 trillion operations per second of AI processing. Running compatible workloads locally reduces dependence on cloud processing, which can cut latency and limit how much sensitive data leaves the device. Cited use cases include real-time translation, AI-assisted content generation, video enhancement, and workflow automation. ASUS cautions that actual benefits depend on the applications, models, and workloads involved.

ASUS frames the QN10 as a platform for agentic AI — systems that execute tasks, coordinate multiple steps, and operate with varying degrees of autonomy. The system supports hybrid workflows that split processing between the local NPU and cloud-based large language model APIs, with the aim of balancing responsiveness, energy consumption, and operating costs. ASUS says the mini PC supports agent frameworks including OpenClaw and Hermes Agent, and users can interact with AI agents through WhatsApp and LINE.

Security stack

The QN10 ships with a multilayered security architecture spanning hardware, firmware, the OS, and cloud-connected environments. It includes firmware-based Trusted Platform Module (fTPM) 2.0 for hardware-based credential protection and Microsoft's Pluton security processor, which strengthens identity protection, safeguards encryption keys, and supports device security and recovery. ASUS notes that organizations still need appropriate software configurations, access controls, and security policies on top of these features.

Developer ecosystem

The system provides access to the Qualcomm AI Hub, which offers more than 175 pre-optimized AI models and supports Bring Your Own Model (BYOM) and Bring Your Own Data (BYOD) workflows. It also works with the Qualcomm AI Stack, Qualcomm's collection of software tools and frameworks for building and running AI applications on its silicon. The platform targets use cases ranging from local large language model inference to edge AI deployments where processing sits closer to where data originates.

Connectivity

Wireless options include Wi-Fi 7 and Bluetooth 6.0. ASUS notes that actual wireless performance depends on network infrastructure, device compatibility, and environmental conditions.

Target buyers

ASUS identifies four groups: creators using AI-assisted editing and content generation tools, developers experimenting with local AI models and agents, enterprises deploying AI-assisted workflows, and professionals in space-constrained environments who want desktop capability without a full tower.

Pricing and availability

The Ascent QN10 reaches US retailers on September 17, 2026 in two configurations:

• 16GB model, $1,349: 16GB LPDDR5x at 8,533MHz, 512GB M.2 NVMe PCIe 4.0 SSD. Sold through Best Buy, Newegg, B&H Photo Video, and the ASUS eShop.
• 32GB model, $1,699: 32GB LPDDR5x at 9,600MHz, 512GB M.2 NVMe PCIe SSD.

Prices and availability vary by retailer.