S Syvidea
Shared platform specification

AMD Strix Halo Local AI Workstation Platform

A compact local AI workstation platform built around AMD Strix Halo, high-bandwidth LPDDR5X memory, dual PCIe 4.0 SSD slots and modern high-speed I/O.

Request a configuration quote → Use recommendation engine →
Syvidea
Memory
128GB
Storage
2TB
AI
50 TOPS
16 cores / 32 threads
Zen 5 cores
128GB
LPDDR5X
2TB
PCIe 4.0 SSD
3
Partner brands
Why a shared platform page

Because the three S1 Series models share the same core spec

This page describes the shared platform that all three S1 Series models are built on. For model-specific differences (chassis, I/O details, dimensions, wireless, thermals, BIOS, packaging, warranty, supply), see each model's detail page or the comparison page.

Syvidea S1 Base

Value option for budget-conscious users and new entrants

S1 Base offers the essential features for local AI at an accessible price point.

View Syvidea S1 Base details →

Syvidea S1 Pro

The recommended choice for most AI builders

S1 Pro is our most popular configuration with proven reliability for daily AI development.

View Syvidea S1 Pro details →

Syvidea S1 Max

Enterprise-grade for professional teams and future scalability

S1 Max is engineered for enterprise reliability and future-proof AI workloads.

View Syvidea S1 Max details →
Specification

Shared platform spec

Identical across S1 Base, S1 Pro, and S1 Max configurations.

Processor & AI

AMD Ryzen AI Max+ 395
Zen 5 • 16 cores / 32 threads • Up to 5.1 GHz
AMD XDNA 2 NPU (50 TOPS)
Up to 126 TOPS total AI performance

Memory & Storage

128GB Unified Memory
LPDDR5X • 256-bit memory interface
2TB SSD (typical configuration)
2× M.2 PCIe 4.0 SSD slots

Graphics & Display

AMD Radeon 8060S Graphics
RDNA 3.5
USB4 / Thunderbolt-class high-speed ports
2.5G Ethernet + WiFi 6/6E

Form Factor & Cooling

Compact desktop workstation
Compact footprint suitable for desk-side or homelab deployment
Active cooling with brand-specific thermal design
External or integrated power supply depending on brand
Why this platform

Designed for local LLM, RAG, agents, and multimodal AI

Large unified memory pool

128GB LPDDR5X with a 256-bit memory interface gives local AI workflows the headroom to run 70B-class quantized models, private RAG contexts, and multi-tool agent stacks on a single machine.

Modern I/O and storage

Dual M.2 PCIe 4.0 SSD slots support fast local model storage and dataset expansion. USB4 / Thunderbolt-class high-speed ports enable modern peripherals and external GPU or storage expansion.

Compact form factor

A compact footprint makes the platform suitable for desk-side deployment, homelab racks, and small-team shared environments where space and noise matter.

NPU + integrated GPU

AMD XDNA 2 NPU and Radeon 8060S integrated graphics provide on-device AI acceleration suitable for inference, multimodal workflows, and supported creative tooling.

What this platform is not for

Honest trade-offs

  • • Not the fastest CUDA image-generation box. If your only goal is maximum Flux or SDXL speed with CUDA / TensorRT, an NVIDIA RTX 5090 workstation is a better fit.
  • • Not for large-scale model training.
  • • Not positioned as a cheap office Mini PC.
  • • Not for users who want direct full-payment online checkout without communication.