Skip to product information
1 of 2
COMING SOONExpected Fall 2026

ASUS ProArt Mini PC NVIDIA RTX Spark AI Desktop

Coming Soon · Expected Fall 2026.

Compact AI Desktop, Powered by the NVIDIA RTX Spark™ Superchip

The ASUS ProArt Mini PC is a NVIDIA RTX Spark™ AI desktop — a compact workstation built on the NVIDIA RTX Spark Superchip, pairing a Blackwell RTX GPU, an NVIDIA Grace (Arm-based) CPU and unified memory in a 150 x 150 mm chassis. Built for AI developers, creators and researchers who want desk-side AI compute to run and fine-tune large models locally — on-device, without the cloud. Coming to Malaysia through EMARQUE — availability to be confirmed.

  • NVIDIA RTX Spark Superchip — up to 1 petaflop of FP4 AI performance from a Blackwell RTX GPU (up to 6,144 CUDA cores) and an NVIDIA Grace Arm-based CPU (up to 20 cores), linked by NVLink-C2C.
  • Up to 128GB unified memory — shared across CPU and GPU for large on-device AI and local LLM workloads.
  • Compact 150 x 150 x 51 mm chassis — with up to 140W of thermal headroom for sustained AI workloads.
  • Built-in connectivity — 10GbE wired networking and an M.2 PCIe Gen 5 x4 slot for fast storage.
  • Desk-side AI compute — local AI development and inference without the cloud.

ASUS ProArt Mini PC Specifications

  • Platform: NVIDIA RTX Spark Superchip (Blackwell RTX GPU + NVIDIA Grace Arm-based CPU)
  • GPU: up to 6,144 Blackwell RTX CUDA cores, 5th-gen Tensor Cores (FP4)
  • CPU: up to 20-core NVIDIA Grace (Arm-based)
  • Unified memory: up to 128GB
  • AI performance: up to 1 petaflop (FP4)
  • Chassis: 150 x 150 x 51 mm
  • Cooling: up to 140W thermal headroom
  • Networking: 10GbE wired
  • Expansion: M.2 PCIe Gen 5 x4
  • Availability: Coming Fall 2026 — Malaysia availability to be confirmed

Specifications, pricing and Malaysia availability are subject to official NVIDIA and OEM confirmation.

Important — pre-launch information

Specifications, pricing and Malaysia availability for NVIDIA RTX Spark products are subject to official confirmation by NVIDIA and the respective OEM. Content on this page reflects information published at announcement. EMARQUE will update each product page as official Malaysia availability and pricing are confirmed.

View full details
  • Up to 1 PFLOPS

    FP4 AI performance per RTX Spark Superchip.

  • Up to 6,144 CUDA cores

    Blackwell RTX GPU architecture.

  • Up to 20-core Grace CPU

    NVIDIA Grace CPU (Arm architecture).

  • Up to 128GB unified

    Single memory pool shared by CPU + GPU.