Preste was chosen by NVIDIA as an embedded software partner (solution advisor - embedded compute); a technical acknowledgement shared with fewer than thirty AI providers in Europe.
Our software expertise enables us to make the most of the power of NVIDIA® Jetson™ hardware platforms (Orin™, AGX Thor™) to build customized, cost-effective embedded AI solutions capable of making real-time decisions at the edge.
Our use cases primarily concern vision – via a Jetson platform, conventional cameras and custom Preste developments, the majority of simple tasks achievable by the human eye and brain can now be automated.
We support you in developing and sizing Jetson-based solutions that optimize technical, functional, and economic factors, in proposing and implementing technical improvements, in optimizing existing solutions.
NVIDIA-dedicated Preste technical stack (indicative): Jetson Orin, Jetson Thor, NVIDIA JetPack™ SDK, Jetson SDK Manager™, TensorRT™, CUDA™ & CUDA toolkit, Deepstream™, Metropolis™ platform, Isaac™ platform, Nvidia cloud GPU (A100, H100, Tesla v100, RTX™).
Accelerated
by
NVIDIA
Jetson
Jetson's AI capabilities
NVIDIA Jetson is the premier AI platform for autonomous machines and embedded applications.
Jetson modules are compact yet high-performance computers, complemented by the NVIDIA JetPack™ SDK for accelerated software development.
Jetson seamlessly integrates with other NVIDIA platforms, sharing AI software and cloud-native workflows while delivering the power-efficient performance required for building software-defined autonomous machines and breakthrough edge-AI solutions.
Benefits
On-site
Operations and decisions are made on-premises, not in the cloud. This is crucial, for example, for activities where internet connectivity is variable.
Real Time
AI-optimized edge architecture, combining
powerful GPUs, dedicated accelerators,
and low latency.
Frugality
Sobriety and performance of embedded models.
Energy efficiency without slowdowns thanks to advanced thermal management.
Governance
Data remains on-site during inference, for complete control over its use and confidentiality.
Real-time applications
Built with Jetson
Performances
Properties | Jetson Orin Nano | Jetson Orin NX | Jetson AGX Thor |
|---|---|---|---|
AI Performance | 20/40 TOPS | 70/100 TOPS | 2070 FP4 TFLOPS |
GPU | 1024-core NVIDIA Ampere GPU with 32 Tensor Cores | 1024-core NVIDIA Ampere GPU with 32 Tensor Cores | 2560-core NVIDIA Blackwell+ 96 Tensor |
CPU | 6-core ARM® Cortex® A78AE v8.2 64-bit CPU | 6-core / 8-core Arm® Cortex® A78AE v8.2 64-bit CPU 1,5 MB / 2MB L2 + 4 MB L3 | 14-core ARM® Neoverse® |
RAM | 4 GB 64-bit LPDDR5 / 8 GB 128-bit LPDDR5 | 8 GB 128-bit LPDDR5 / 16 GB 128-bit LPDDR5 | 128GB, 273 GB/S |
Storage | NVMe (128GB/ 256GB / 512 GB / 1TB) | NVMe (128GB/ 256GB / 512 GB / 1TB) | NVMe (128GB/ 256GB / 512 GB / 1TB) |
Power consumption | 7W - 15W | 10W - 40W | 40-130W |
