Specifications
Manufacturer
Inventec
Platform
NVIDIA IGX Thor
AI Compute
Up to 5,581 TFLOPS (Blackwell)
Workloads
Generative AI at the edge —
VLM & VLA models
Safety
Functional Safety Island
(inside-out) + NVIDIA Halos (outside-in)
Role
Secure command tower for
human-robot collaboration
Announced
COMPUTEX 2026 (Taipei)
Target
Factories, autonomous robots,
industrial safety & throughput
Overview
The Inventec Atlas Edge AI Server is a physical AI solution built on NVIDIA IGX Thor, delivering up to 5,581 TFLOPS of Blackwell-powered compute to run generative AI models — including vision-language models (VLM) and vision-language-action models (VLA) — directly at the edge.
Designed for human-robot collaboration, the Atlas combines a Functional Safety Island for inside-out safety with the NVIDIA Halos blueprint for outside-in safety, enabling complete robot perception. It serves as a secure command tower that lets autonomous robots operate safely alongside human workers while increasing throughput and compliance in factories.
Announced at COMPUTEX 2026 as the centerpiece of Inventec's AI Factory infrastructure domain, the Atlas exemplifies how the NVIDIA ecosystem is being deployed to commercialize physical AI across manufacturing and logistics.
Request a Quote — Inventec Atlas Edge AI Server
QS Compute — global B2B supply of AI computing hardware. Price on request, stock availability confirmed on inquiry.
Get Your Quote →