Overview
Second-generation TT-QuietBox desktop AI workstation powered by 2× Tenstorrent Blackhole p300c processors, liquid-cooled for whisper-quiet continuous operation. Capable of running models up to 120B parameters on-premise. Ships with a fully open-source software stack (TT-Metal) supporting PyTorch and ONNX workflows, starting at $9,999 — $2,000 less than the original TT-QuietBox.
Key Features
- 2× Tenstorrent Blackhole p300c processors for teraflop-class AI inference
- Liquid-cooled design for sustained silent operation under continuous workloads
- Runs models up to 120B parameters locally without cloud dependency
- Fully open-source software stack (TT-Metal) — no vendor lock-in
- Compact tower form factor, redesigned lighter and more durable than TT-QuietBox v1
Ideal For
AI developers and research teams needing on-premise 120B parameter inference with full software stack visibility and no GPU vendor lock-in.
AI Processors
2× Tenstorrent Blackhole p300c
Cooling
Liquid cooling
Architecture
RISC-V (Tensix Cores)
Software Stack
Open-source TT-Metal (ONNX, PyTorch)
Max Model Size
Up to 120B parameter models
Availability
Ships in 6-12 weeks
$9,999.00
Prices may vary. Verify on vendor site.
Quick Specs
- AI Processors
- 2× Tenstorrent Blackhole p300c
- Cooling
- Liquid cooling
- Architecture
- RISC-V (Tensix Cores)
- Software Stack
- Open-source TT-Metal (ONNX, PyTorch)
- Max Model Size
- Up to 120B parameter models
- Availability
- Ships in 6-12 weeks
Tags
llm-inferencellm-trainingdata-analytics
