Jetson T2000, NVIDIA Embedded AI Module with Future compact Jetson Thor module for low-power edge AI

NVIDIA Jetson T2000 is a planned entry-level Jetson Thor module for compact robotics and edge AI. The supplied planning file identifies a Blackwell GPU, a Neoverse CPU and a low-power positioning, with availability indicated for 2027 Q1; final memory, AI-performance and ordering details must be confirmed before publication.

Best Fit Applications

Mainstream Robotics

The planning profile targets robotics products that need a lower-cost entry into the Jetson Thor physical-AI platform.

Compact Edge Vision

The planned low-power positioning is suited to compact vision systems and intelligent devices, subject to final specifications.

Education and Prototyping

The entry positioning may suit teaching and early development, but final availability and kit options must be confirmed.

Small Equipment Makers

The product is positioned for smaller manufacturers that want to explore Thor-based edge AI without the highest compute tier.

Why Jetson T2000

Entry Point to Jetson Thor

The supplied planning data positions T2000 as a compact, lower-power route into the next-generation Thor ecosystem.

Broader Device Deployment

The planned low-power profile is intended to make physical AI more practical for mainstream robotics and compact devices.

Validate Before Ordering

Because final memory and performance figures are not present in the planning data, product specifications must be confirmed before publication.

Compare Related Models

	NVIDIA Jetson T2000	NVIDIA Jetson T3000	NVIDIA Jetson T4000	NVIDIA Jetson Orin Nano 8GB
AI Performance	Entry-level Thor performance; final figure to be confirmed	865 FP4 TFLOPS, according to supplied planning data	Up to 1,200 FP4 TFLOPS, sparse	Up to 67 TOPS
GPU	Blackwell GPU; final core count to be confirmed	Blackwell GPU; final core count to be confirmed	1,536 CUDA cores; 5th-generation Tensor Cores	1,024 CUDA / 32 Tensor Cores
CPU	Neoverse CPU; final core count to be confirmed	8-core Neoverse Arm CPU	12-core Arm Neoverse V3AE	6-core Arm Cortex-A78AE
Memory	Final memory configuration to be confirmed	32GB LPDDR5X	64GB LPDDR5X	8GB LPDDR5
Power	Low-power profile according to supplied planning data	Approximately half of the T5000 power profile; confirm final value	40–70W	7–25W
MIG	Thor module feature set to be confirmed	Thor module feature set to be confirmed	6 TPC MIG	Jetson Orin platform
Best Fit Scenario	Mainstream robotics, compact edge vision and entry physical-AI projects	Scaled humanoid robotics, mainstream robots and edge AI deployments	Production robots, industrial vision and edge servers with a balanced Thor profile	Entry robotics, vision AI and education
Target Customer	Education teams, small equipment makers and entry robotics developers	Robot mass-production teams and embedded device manufacturers	Robot makers, industrial automation teams and edge-compute integrators	Makers, educators and small device makers
Not Recommended For	Production designs that require confirmed memory, power and performance specifications	Fully specified production deployments before final NVIDIA ordering data is available	Large physical-AI workloads that need the T5000 memory and compute tier	Large physical-AI workloads