
NVIDIA Orin Nano Super
Compact computing for onboard processing, sensors and autonomy.
- Weight
- 500 g
- Power input · up to
- 25 WPoE++
- Enclosure
- 120 × 120 × 40 mm
Blackhole
Blackhole puts NVIDIA computing into rugged hardware for vehicles, autonomous platforms and remote installations.
Compact onboard processing or larger local AI workloads. Choose the compute platform around the work and the installation.

Compact computing for onboard processing, sensors and autonomy.

Grace Blackwell computing for local AI development and inference.

Blackhole Orin Nano’s sealed housing conducts heat from the compute platform to its exterior, bringing protection and cooling into the same physical assembly.
Orin Nano’s exterior fins increase the housing’s heat-transfer area. Blackhole GB10 combines its finned enclosure with internal blowers and a heat-exchanger assembly.
Custom power and data PCBs connect the compute platform, input power and interfaces inside the rugged assembly.
Mounting features and accessible connectors support installation and servicing.


| Specification | NVIDIA Orin Nano Super | GB10 Grace Blackwell |
|---|---|---|
| CPU | 6-core Arm Cortex-A78AE | 20-core Arm · 10 Cortex-X925 + 10 Cortex-A725 |
| GPU | Ampere · 1,024 CUDA cores · 32 Tensor cores | Blackwell |
| Memory | 8 GB LPDDR5 · 102 GB/s | 128 GB unified LPDDR5x |
| Peak AI compute | Up to 67 INT8 TOPS | Up to 1 PFLOP · sparse FP4 |
| Storage | microSD slot and NVMe SSD support | 2 TB M.2 2230 QLC PCIe Gen4 SSD |
| Specification | Blackhole Orin Nano | Blackhole GB10 |
|---|---|---|
| Dimensions | 120 × 120 × 40 mm | 188 × 201 × 55 mm |
| Weight | 500 g | 1,500 g |
| Power input | PoE++ · up to 25 W | USB-C Power Delivery · up to 280 W |
| Construction | Sealed housing with integrated heatsink; custom power and data PCBs. | Finned enclosure with two internal blowers and a heat exchanger. |
| Interfaces | Ethernet / PoE, USB-C, fiber; optional M.2 wireless. | USB-C ×4, HDMI, 10GbE RJ45, QSFP ×2. |

Choose Blackhole as a preconfigured Quantum Helix appliance, or use the hardware with another workload.
Quantum Helix provides workload identity and controlled connections between participating applications.
Explore Quantum HelixTell us about your workload, power budget and installation. We’ll discuss the Blackhole configuration and qualification requirements that fit.