← All hardware
Physical-AI SoC running GenAI at under 10W, 50-200 TOPS
Pros
- GenAI at very low power
- Full SoC (CPU+vision+AI)
- Scalable performance
- Industrial-grade options
Cons
- Inference/edge-focused
- Smaller ecosystem
- GA for PCIe card mid-2026
- Less brand recognition
✓ Where it shines / best for
- Robotics / physical AI
- Industrial edge servers
- Multimodal edge inference
- Low-power GenAI
✕ Not the best fit for
- Datacenter training
- CUDA workflows
- Large-cluster scale
- Consumer plug-and-play
Features
- ✓ 6nm MLSoC for physical AI
- ✓ Runs GenAI (LLM/LMM) <10W
- ✓ 50-200 TOPS scalable
- ✓ Arm Cortex-A65 + ARC EV74
- ✓ PCIe Gen5, 10GbE
- ✓ Modalix PCIe HHHL card (8/16GB)
Pricing
| Plan | Price | Billing | Notes |
|---|---|---|---|
| Modalix PCIe card | Contact (sampling) | one-time | GA Q2 2026; 8GB/16GB variants |
Pricing verified from the official source. Prices change often — confirm on the vendor's site before buying.
Specifications
| use | Edge/physical AI inference |
| power | <10W (SoC) to card-level |
| memory | 8GB or 16GB (PCIe card) |
| performance | 50-200 TOPS |
| architecture | MLSoC Modalix (6nm) |
Sponsored
A full review is being generated for this product and will appear here shortly.