Accelerate AI Inference at the Edge with NVIDIA Jetson T4000

NVIDIA has recently announced NVIDIA Jetson™ T4000. This new module delivers up to 1200 FP4 TFLOPs of AI compute and 64 GB of memory. Thanks to its capabilities, NVIDIA Jetson™ T4000 provides an ideal balance of performance, efficiency, and scalability, which helps bring high-performance AI and real-time reasoning to a wider range of robotics and edge AI applications. 

The new NVIDIA Jetson™ T4000 module has 1× NVENC and 1× NVDEC hardware video codec engines to enable real-time 4K video encoding and decoding. This design makes it possible to use the module to combine advanced vision processing and I/O capabilities with power and thermal efficiency. The Jetson T4000 module shares the same form factor and pin compatibility with the NVIDIA Jetson T5000 module. Developers can design common carrier boards for both modules, taking the differences in thermal and other inherent module features into account. Table below highlights the differences between the modules.

Both modules deliver strong performance for a number of text-to-speech (TTS), large language models (LLMs) and vision-language-action (VLA) models. The new Jetson T4000 delivers up to 2x performance compared to previous generations. The table below displays performance metrics of T4000 and T5000.

FORECR’s DSBOARD-THRMAX carrier board and DSBOX-THRMAX industrial edge AI computer were initially introduced with support for Jetson™ T5000, delivering high-performance AI computing based on NVIDIA Blackwell architecture for demanding robotics, autonomous systems, and industrial AI workloads. Now, with the addition of Jetson™ T4000 support, these platforms address a broader set of use cases within a lower power envelope, ranging from robotics to edge AI applications.

The DSBOARD-THRMAX carrier board offers a compact 140 ×125 mm form factor with high-speed expansion, including PCIe Gen5storage, QSFP+ (3×25GbE) networking, dual HDMI 2.1, and abroad set of industrial interfaces such as CAN bus, RS-232/422/485, and digital I/Os. Designed for flexible system integration, it supports M.2 expansion, wireless connectivity (Wi-Fi, Bluetooth, LTE/5G), and high-bandwidth camera configurations.

The DSBOX-THRMAX industrial edge AI computer extends these capabilities into a rugged, enclosed system designed for mission-critical deployments. Thanks to its extended operating temperature range and rich I/O—including high-speed Ethernet, USB-C, PCIe Gen5 storage, and optional support for up to eight GMSL2 cameras—the DSBOX-THRMAX is ideal for sensor-rich robotics, vision-based AI systems and autonomous platforms.