Apple-Compete AI Mini PC

Hala Core (STHT1)

Desktop Power,
Palm-Sized Design

The Hala Core fits a 54W AMD FP8-H processor, USB4 V2 at 80 Gbps, 10-gigabit Ethernet, and a built-in stereo speaker system into a premium 127mm metal chassis weighing just 450g. Enterprise-grade local AI in the footprint of a coaster.

AMD FP8-H 54W processor
USB4 V2 80Gbps connectivity
Built-in stereo speakers + 4-mic
Hala Core Mini PC front perspective
0WTDP
0GbpsUSB4 V2 Speed
0GEthernet
0gWeight
Features

Engineered To Compete With The Best

Premium connectivity and built-in audio in a chassis smaller than a paperback.

icon

AMD FP8-H Processor

54W TDP with vapor chamber thermal module. Handles AI inference, multi-threaded workloads, and always-on edge deployments with ease.

icon

USB4 V2 at 80 Gbps

Dual USB4 V2 ports with DisplayPort alt mode. Transfer massive datasets and drive external displays at blazing speed.

icon

10-Gigabit Ethernet

Realtek RTL8127AT 10G LAN for data-center-grade network throughput. Ideal for NAS, inference clusters, and high-bandwidth edge nodes.

icon

Built-In Audio System

SN6242 Soundwire codec, dual stereo speakers, and optional 4-mic array. Conference calls, voice AI, and media without external hardware.

Hala Core front panel showing USB-C ports

127mm. 450 Grams.
Zero Compromise.

The Hala Core packs a 54W AMD processor, dual USB4 V2 at 80 Gbps, 10-gigabit Ethernet, built-in stereo speakers, and a quad-mic array into a premium metal chassis that fits in the palm of your hand. Add dual HDMI 2.1 outputs for 4K@120Hz displays and WiFi 7 — all in a device that weighs less than a pound.

Hala Core rear panel showing all ports

🚀Powerful AI Processing

  • AMD FP8-H processor — 54W TDP with vapor chamber cooling
  • Integrated RDNA graphics for GPU-accelerated inference
  • LPDDR5 quad-channel memory (4× 315-ball modules)

🔗Premium Connectivity

  • 2× USB4 V2 at 80 Gbps with DisplayPort alt mode
  • 2× HDMI 2.1 TMDS — 4K@60Hz, 4K@120Hz with DSC
  • 10G LAN (Realtek RTL8127AT) + WiFi 7 + Bluetooth 5+

Compact Metal Design

  • 127×127×50mm — smaller than an Apple Mac Mini
  • ~450g all-metal body with plastic bottom cover
  • 2× M.2 2280 PCIe 4.0×4 NVMe SSD slots

Endless Possibilities With The Hala Core

Hala Core bottom perspective

Desk-Side AI Companion

Drop it next to your monitor for always-on local LLM inference, RAG pipelines, and voice AI — with zero cloud latency.

Conference Room Hub

Built-in speakers and quad-mic array make it a standalone video-conferencing endpoint with on-device AI transcription.

Edge Network Node

10G LAN and USB4 V2 turn the Hala Core into a high-throughput edge gateway for real-time analytics and sensor fusion.

Silent Media Server

Stream 4K@120Hz to dual HDMI 2.1 displays while serving media across your network — all in near-silence.

One Device. Every
Deployment.

From the conference room to the server closet, the AKB56 adapts to your workload. Use its built-in audio for smart office AI, leverage 10G LAN for high-throughput edge computing, or deploy fleets with enterprise-grade security — all in a silent 450g metal chassis.

icon

Local AI & Inference

Run LLMs, RAG pipelines, and voice AI on-device with 54W sustained compute and massive memory bandwidth.

  • Local LLM Hosting
  • Voice Agents
  • RAG Pipelines
icon
icon

Smart Office & Conferencing

Quad mic + stereo speakers for AI-powered meeting rooms, transcription, and voice assistants — no external audio needed.

  • AI Transcription
  • Video Conferencing
  • Room Automation
icon
icon

High-Speed Networking

10G LAN + WiFi 7 for NAS gateways, inference clusters, and bandwidth-intensive edge deployments.

  • NAS Gateway
  • Cluster Node
  • Data Ingest
icon
icon

Enterprise Endpoints

TPM (SPI), Modern Standby option, Windows 11 & Linux for managed fleet deployments in regulated environments.

  • Fleet Management
  • Data Sovereignty
  • Kiosk / POS
icon

Key Specifications

Everything packed into 127mm of precision-engineered metal

Processor

AMD FP8-H

54W TDP with vapor chamber thermal solution. Runs Windows 11 and Linux for maximum deployment flexibility across AI inference, edge compute, and office workloads.

Compatibility: Supports: Llama, DeepSeek, Qwen, Mistral, Whisper, and all major OSS models

Graphics

Integrated

AMD RDNA

Memory

LPDDR5

4-ch (315-ball)

Storage

2× M.2 2280

PCIe 4.0×4 NVMe

Rear I/O

USB4 V2

80Gbps + HDMI 2.1

Networking

WiFi 7

10G LAN · BT 5+

Audio

SN6242

Speakers + 4-Mic

Display

2× HDMI 2.1

4K@120Hz DSC

Chassis

127mm

127×127×50mm · 450g

Ready To Deploy The Hala Core?

80Gbps USB4 V210G EthernetBuilt-in speakers
Hala Core view 1
Hala Core view 2
Hala Core view 3