Wolf Advanced Technology (WOLF) has developed a new range of rugged embedded system modules powered by the NVIDIA® Jetson Thor™ platform, marking a significant advancement in edge AI computing capabilities for defense and aerospace applications.
Designed to support next-generation autonomous systems, generative AI workloads, and real-time multi-sensor processing, the Jetson Thor platform delivers unmatched AI performance and configurability in a compact form factor.
Greg Maynard, CTO at Wolf Advanced Technology, said, “NVIDIA Jetson Thor modules unlock new possibilities for intelligent machines and embedded edge AI. We’re thrilled to be working on products that will harness this power, and we invite customers and partners to engage with us early to shape solutions that fit their unique needs.”
Built on NVIDIA’s latest Blackwell architecture, Jetson Thor offers up to 2070 FP4 TFLOPS of AI compute performance and integrates 128 GB of high-bandwidth LPDDR5X memory. Its modular design supports power configurations ranging from 40 W to 130 W, enabling scalable performance to suit a wide range of mission requirements.
Jetson Thor introduces Multi-Instance GPU (MIG) support to the Jetson family for the first time. MIG provides secure hardware-level partitioning of GPU resources, ensuring predictable latency, throughput, and quality of service. This enables reliable workload isolation, a key requirement for safety-critical and mission-assured operations.
WOLF’s new Jetson Thor-based modules are engineered to deliver low-latency, real-time processing for demanding multi-sensor applications. By combining Thor’s processing power with WOLF’s advanced networking and I/O architecture, these modules are optimized for operation in rugged and deployed environments.
Key enhancements include:
- ConnectX-7 integration for expanded high-speed networking options, including PCIe Gen5 and 200GigE capabilities.
- Support for PCIe Gen4 x16, 100G, and dual 25G configurations
- 14-core Arm® Neoverse®-V3AE CPU, third-generation Programmable Vision Accelerator (PVA), dual video encoders/decoders, and an optical flow engine
These enhancements significantly increase throughput and bandwidth beyond Jetson Thor’s baseline specifications, making the solution suitable for data-intensive ISR and sensor fusion workloads.
WOLF’s integration of its FGX technology enables seamless support for essential video and sensor interfaces such as ARINC-818, CoaXPress, SDI, and analog video. These features are critical for high-performance embedded systems in aerospace and defense platforms.
In addition, WOLF supports IEEE 802.1DP TSN (Time-Sensitive Networking) for deterministic, real-time data transfer, and includes ≥1TB of secure onboard storage to handle high-throughput AI and signal processing applications.
Building on the success of its SOSA® Aligned NVIDIA Orin-based solutions, WOLF plans to launch flagship 3U and 6U VPX modules aligned with the SOSA Technical Standard in early 2026. The company’s agile engineering approach and modular design practices also enable rapid customization and integration, allowing system designers to quickly deploy Jetson Thor into existing or new rugged platforms.
