YUAN Expands the NVIDIA Jetson Thor Portfolio with New T3000 and T2000 Modules
Next-Generation Edge AI Platforms from YUAN Accelerate Mainstream Robotics and Scalable Physical AI Deployments Built on NVIDIA Blackwell Architecture
YUAN High-Tech (YUAN), an NVIDIA ecosystem partner, today announced it has begun development of its next-generation edge AI platforms based on the new NVIDIA Jetson T3000 and T2000 modules. Built on the same NVIDIA Blackwell architecture and software stack as the flagship NVIDIA Jetson AGX Thor T5000, these new modules deliver a seamless, low-friction migration path for future robotics and edge AI applications while providing unparalleled flexibility for scalable deployments.
Powered by the flagship Jetson AGX Thor T5000, YUAN has successfully deployed production-ready edge AI platforms for factory inspection, autonomous robotics, and intelligent safeguarding, delivering low-latency edge AI while keeping data securely processed on-site. As robotics enters the era of large-scale commercialization, developers increasingly require AI platforms that balance extreme performance, power efficiency, compact size, and deployment cost.
Jetson T3000: High Performance for Mainstream Robotics
Compared with the flagship Jetson AGX Thor T5000, the new Jetson T3000 reduces memory capacity, power consumption, and module size by approximately 50%, while preserving the same NVIDIA software stack and delivering around 90% of the T5000's AI inference performance. This enables developers to significantly reduce system footprint, power consumption, and overall deployment costs without sacrificing AI capability, making the T3000 an ideal platform for scalable robotics and edge AI. In addition, the IGX T3000 retains the Functional Safety Island (FSI), providing the essential functional safety required for industrial robots, autonomous systems, and safety-critical edge AI applications.
Jetson T2000: Bringing Thor Architecture to More Edge AI Applications
The new Jetson T2000 extends the powerful Thor architecture to a broader spectrum of edge AI systems. Featuring 400 FP4 TFLOPS of AI compute, a NVIDIA Blackwell GPU, 16 GB LPDDR5X memory, and a power envelope of only 40 W, it offers an exceptionally attractive and cost-effective entry point for visual AI agents, autonomous mobile robots (AMRs), robotic arms, and other cost-sensitive intelligent edge applications.
Immediate Development Path via Emulation
Although Jetson T3000 and T2000 modules are expected to be available in Q1 2027, development can begin using T3000 emulation mode available later this month with JetPack 7.2.1. . Utilizing YUAN's existing development platform powered by Jetson AGX Thor, customers will be able to leverage T3000 Emulation Mode to validate applications using the complete NVIDIA software stack, including open models such as NVIDIA Nemotron, NVIDIA Cosmos 3, NVIDIA Isaac GR00T, and the latest Jetson AI Agent capabilities. T2000 emulation support is planned for a future release.
One Thor Platform. Every Deployment Scale.
From the flagship Jetson T5000 to the highly efficient T3000 and the cost-optimized T2000, YUAN provides a comprehensive Jetson Thor platform portfolio. This allows customers to smoothly migrate existing applications while accelerating next-generation robotics products from development to high-volume production.