Advantech Expands Its Edge AI Platform Portfolio Powered by New NVIDIA Jetson T2000 and T3000 Modules for the Next Generation of Physical AI
2026/07/16
TAIPEI, Taiwan - July 16, 2026 - Advantech (TWSE: 2395), a global leader in industrial IoT, has announced its edge AI portfolio powered by NVIDIA Jetson T2000 and T3000 modules, providing customers with a complete range of AI computing platforms from entry level commercial systems to flagship industrial solutions.
As AI enters the era of Physical AI, robots and autonomous machines require increasingly powerful, energy efficient, and compact AI computing to support workloads such as large language models, vision language models, and AI agents at the edge. By supporting the new NVIDIA Jetson T2000 and T3000 alongside existing T4000 and T5000 based platforms, Advantech enables customers to scale AI deployments according to their performance, power, and cost requirements while accelerating innovation across robotics, autonomous machines, and intelligent edge applications.
NVIDIA Jetson T2000 and T3000 Deliver AI performance in smaller, lower-power, lower-cost module
At the forefront of this revolution is NVIDIA Jetson Thor, a new class of robotics supercomputer designed to power the next generation of humanoid robots and autonomous machines. As humanoid robotics move toward high-volume production and mass-market adoption, the industry needs a compact, power-efficient AI supercomputer optimized for scalable deployment.
Designed for mainstream robotics and edge AI deployments, NVIDIA Jetson T3000 delivers up to 865 FP4 TFLOPS of AI performance with 32GB LPDDR5X memory at just 70W, providing similar inference performance of NVIDIA Jetson T5000 while reducing power consumption and memory costs, which making it ideal for scalable robotics deployments where performance and power efficiency are critical.
NVIDIA Jetson T2000 targets a broader edge AI applications, offering 400 FP4 TFLOPS of AI performance with 16GB LPDDR5X memory within a 40W power envelope. It's the ideal entry point to the Thor product family, enabling cost-effective deployment of AI-powered machines at scale.
Together, the new modules extend the NVIDIA Jetson Thor portfolio with optimized options that balance AI performance, energy efficiency, compact form factors, and affordability for a broader range of edge AI deployments.
Developers can accelerate the development of next-generation Jetson Thor modules with software emulation and agent skills for memory optimization.
Advantech Delivers a Complete NVIDIA Jetson Thor Platform Portfolio
Building on the expanded NVIDIA Jetson Thor family, Advantech is extending its Edge AI platform portfolio to support the complete Jetson Thor series, including Jetson T2000, T3000, T4000, and T5000 modules.
MIC-AI and MIB Series
Purpose-built for industrial Physical AI deployment, Advantech's MIC-AI systems, powered by NVIDIA Jetson Thor, combine next-generation AI performance with industrial-grade reliability for robotics, transportation, and industrial automation. The portfolio enables customers to accelerate edge AI deployment with long lifecycle support, full-stack NVIDIA software, flexible system integration, and proven industrial design for demanding real-world environments.
The MIC-743-AT delivers the AI performance required for VLMs, LLMs, and real-time AI inference, while the MIC-742-AT provides a compact and flexible platform optimized for robotics and Physical AI applications.
For customers seeking flexible integration options, Advantech also offers AI boards powered by NVIDIA Jetson Thor, including the MIB-741-AT and MIB-742-AT. Designed as the AI computing core for robotics and autonomous systems, the MIB series delivers high-bandwidth sensor integration capabilities and support for NVIDIA JetPack, NVIDIA CUDA-X libraries, NVIDIA TensorRT, and NVIDIA Holoscan, enabling real-time perception, reasoning, and action for Physical AI applications, including humanoid robots, mobile manipulators, and industrial autonomous systems.
AFE Series and AIR Series
Purpose-built for advanced robotics applications, the Advantech AFE-A702 robotic controller system and ASR-A702 robotic controller board harness the full spectrum of NVIDIA Jetson Thor performance to enable transformer-based reasoning, multimodal AI inference, and real-time multi-sensor perception for humanoids, AMRs, and unmanned vehicles. With support for multiple cameras, LiDAR, IMUs, NVIDIA JetPack 7.2, NVIDIA Isaac ROS, NVIDIA Isaac Sim, Holoscan, and Advantech Robotic Suite, these platforms streamline robotics development, system integration, and deployment. Combined with NVIDIA NemoClaw and Jetson agent skills, they extend edge AI agents into physical robotic systems, providing a scalable foundation for industrial assistants and next-generation autonomous applications.
Advantech AIR-075 is purpose-built for multi-camera vision and agentic AI applications. Its multiple native 10GbE RJ45 interfaces provide the high bandwidth required to connect and process multiple IP camera streams in real time, eliminating the need for external network aggregation switches. This direct-connect architecture reduces system complexity and integrates seamlessly with existing IP-based factory vision infrastructure. Supported by Advantech’s edge AI software ecosystem, NVIDIA JetPack 7.2, and NVIDIA open models and frameworks, including NVIDIA Cosmos, NVIDIA Nemotron, and NVIDIA Isaac GR00T, AIR-075 helps simplify AI development, deployment, and fleet-scale operations.
Advantech and NVIDIA Empower the Next Generation of Physical AI Applications
Advantech plans to introduce Edge AI platforms supporting the NVIDIA Jetson T2000 and T3000 modules following NVIDIA’s module availability, with commercial availability expected in the first quarter of 2027. As an NVIDIA Elite Partner, Advantech works closely with NVIDIA to accelerate the adoption of Physical AI by delivering scalable, flexible, and industrial-ready Edge AI platforms. Through its comprehensive NVIDIA Jetson Thor portfolio, Advantech enables developers and enterprises to bring next-generation robotics, autonomous machines, and intelligent edge applications from concept to real-world deployment faster.
-25/產品排排站-(12)20251229111138.png)
-25/MIB-73520260312212621.png)
-25/ASR-A702-top-side-for-PIS_compressed20260611095432.png)
-25/產品排排站-(11)20251229110635.png)
-25/產品排排站-(8)20251017153609.png)
