LOCAL AI APPLIANCE

3A Mate

Private on-prem AI compute for agents, robots and enterprise workflows. Powered by NVIDIA Jetson T5000.

$7,999.00

Pre-installed agent 3A.ai Agent Hermes Agent

Local LLM Qwen 3.6 pre-installed No pre-installed LLM

Ships pre-configured · Questions? Call sales at (866) 236-3835

Private on-prem AI compute for agents, robots, and enterprise workflows. 3A Mate is a complete AI appliance powered by NVIDIA Jetson T5000. Run open-source models entirely on-device: no cloud, no subscriptions, no per-token costs.

Plug & play. Arrives pre-configured with your selected agent and model. Power on, connect, and start inferencing within minutes.
Local & private. Designed to run fully offline in secure environments.
Open & extensible. Standard APIs, open-source model support and a full Linux environment (Ubuntu 22.04 LTS + NVIDIA JetPack).

Choose your pre-installed agent (3A.ai Agent or the open-source Hermes Agent) and whether to pre-install Qwen 3.6.

Why 3A Mate

A complete AI appliance, not a dev kit.

Private on-prem compute on your desk, with total control over your models and data.

Local AI, zero cloud

Run open-source LLMs entirely on-device. No internet required, no per-token costs, no rate limits.

Keep data on-site

Models run locally, so prompts and documents are processed on the device in your own facility.

Plug & play

Arrives pre-configured with your selected models. Power on, connect and start inferencing within minutes.

Jetson T5000 powered

NVIDIA Blackwell architecture with up to 2070 FP4 TFLOPS and 128GB LPDDR5X unified memory.

Air-gapped ready

Designed to operate fully offline in secure environments.

Open & extensible

Standard APIs, open-source model support and a full Linux environment. Customize and extend without lock-in.

Specifications

Technical specifications.

Compute

AI compute (FP4): Up to 2070 FP4 TFLOPS
GPU: NVIDIA Jetson T5000 (Blackwell)
CUDA cores: 2048
Tensor cores: 64 (4th Gen)
CPU: 16-core NVIDIA Arm Cortex-A78AE

Memory & storage

Memory: 128GB LPDDR5X (unified)
Memory bandwidth: 273 GB/s
Storage: 1TB NVMe SSD (expandable)

Connectivity

Networking: 10GbE Ethernet
PCIe: PCIe Gen4 x8
USB: USB 3.2 (x4), USB 2.0 (x2)
Display: HDMI 2.1, DP 1.4a

Physical & power

Power consumption: 100W (typical)
Power supply: 150W external adapter
Form factor: Compact desktop (approx. 5L)
Operating system: Ubuntu 22.04 LTS + NVIDIA JetPack

FAQ

3A Mate questions.

Contact sales

What is 3A Mate?

3A Mate is a complete local AI appliance powered by NVIDIA Jetson T5000 with 128GB of unified memory. It runs open-source models on-device for agents, robots and enterprise workflows.

Does 3A Mate need the cloud or an internet connection?

No. Models run on the device and 3A Mate is designed to operate fully offline, so there are no per-token costs or rate limits.

Which agent and model come pre-installed?

Choose the 3A.ai Agent (3A's own agent platform) or the open-source Hermes Agent from Nous Research, and choose whether Qwen 3.6 is pre-installed and ready to run on first boot.

Can I buy for my university, integrator or company?

Yes. Contact our sales team for volume quotes and to discuss your deployment.

Get started