AIToday

Aetina adds Nvidia Jetson T3000, T2000 support for edge robotics

Robotics & Automation News4h ago
Aetina adds Nvidia Jetson T3000, T2000 support for edge robotics

Key takeaway

Edge AI computing specialist Aetina announced support for Nvidia's new Jetson T3000 and T2000 modules in its upcoming hardware platforms designed for robotics and industrial edge AI. The T3000 delivers up to 865 FP4 TFLOPS at 70 W power consumption, while the lower-power T2000 offers up to 400 FP4 TFLOPS, both running on Nvidia Blackwell GPUs. Aetina, an Nvidia Elite Partner, will offer these modules in both fan-based and fanless variants to scale robotics deployments from development to production, with the modules expected to launch in the first quarter of 2027.

Summaries like this, in your inbox every morning.

Sign up free →

3 Key Points

  • What happened

    Edge AI specialist Aetina announced it will support Nvidia's newly introduced Jetson T3000 and T2000 modules in its upcoming DeviceEdge AIE-KT fan-based and AIE-PT fanless systems, targeting robotics and industrial edge AI applications.

  • Why it matters

    The T3000 delivers up to 865 FP4 TFLOPS at 70 W power, while the T2000 offers up to 400 FP4 TFLOPS, giving developers scalable options for robotics projects from development through production without sacrificing power efficiency—a key constraint in robotics and edge deployment.

  • What to watch

    Nvidia expects both modules to become available in the first quarter of 2027; Aetina plans to announce additional details on system availability, specifications and developer support closer to launch.

In Depth

Aetina, an Nvidia Elite Partner specializing in edge AI computing, announced plans to integrate support for Nvidia's new Jetson T3000 and T2000 modules into two upcoming hardware families: the DeviceEdge AIE-KT series and the DeviceEdge AIE-PT series. The move positions Aetina among the first launch partners for these modules and reflects the company's strategy to offer scalable AI computing options for robotics and industrial edge applications.

The Jetson T3000 module delivers up to 865 FP4 TFLOPS of AI computing performance using an Nvidia Blackwell GPU with 1,536 CUDA cores, an eight-core Arm Neoverse CPU, and 32 GB of LPDDR5X memory, while operating at up to 70 W. The lower-power T2000 offers up to 400 FP4 TFLOPS with an Nvidia Blackwell GPU featuring 1,024 CUDA cores and 16 GB of LPDDR5X memory, targeting applications such as visual AI agents, autonomous mobile robots, and robotic manipulators. Both modules will support Nvidia's AI software ecosystem, including Nvidia Nemotron, Nvidia Cosmos 3, Nvidia GR00T, and Jetson agent skills.

Aetina's two product lines address different deployment scenarios. The DeviceEdge AIE-KT is a fan-based platform designed for applications requiring sustained performance and additional thermal capacity, while the new DeviceEdge AIE-PT series provides a fanless design for enclosed, rugged, and space-constrained environments where silent operation and reduced maintenance are priorities. According to Aetina, these product lines will allow customers to scale robotics projects from development through production deployments. For context, Aetina's existing systems based on the Nvidia Jetson AGX Thor platform deliver up to 2,070 FP4 TFLOPS of AI computing performance within a 130 W power envelope and are capable of running generative AI models like Nvidia Isaac GR00T with low latency for humanoid robots and other physical AI applications.

Nvidia expects the Jetson T3000 and T2000 modules to become available during the first quarter of 2027. Aetina, as one of the first launch partners, has stated it will announce additional details on system availability, specifications, and developer support closer to the launch date.

Context & Analysis

Aetina's announcement marks the company's continued expansion as an Nvidia Elite Partner into lower-power edge AI modules. The company's existing Jetson AGX Thor systems already deliver up to 2,070 FP4 TFLOPS within a 130 W power envelope and can run generative AI models like Nvidia Isaac GR00T for humanoid robots and physical AI applications. The new T3000 and T2000 modules represent a step down in absolute performance but offer more granular power-efficiency options: the T3000 at 865 TFLOPS and 70 W sits between the Thor platform and lighter workloads, while the T2000 at 400 TFLOPS addresses space-constrained and silent-operation use cases that the higher-power systems cannot serve.

By introducing both fan-based (AIE-KT) and fanless (AIE-PT) hardware families around these modules, Aetina is addressing the full deployment lifecycle: fan-based systems for sustained performance in less restrictive environments, and fanless platforms for rugged, enclosed, or silent-operation scenarios common in production robotics. This modular approach allows customers to select the thermal design and performance tier that best matches their robotics project phase and operating environment, from prototyping through scaled production deployment.

FAQ

When will the Nvidia Jetson T3000 and T2000 become available?
Nvidia expects both modules to become available during the first quarter of 2027.
What are the performance and power specs of each module?
The T3000 delivers up to 865 FP4 TFLOPS with an Nvidia Blackwell GPU, 1,536 CUDA cores, an eight-core Arm Neoverse CPU, and 32 GB of LPDDR5X memory, operating at up to 70 W. The T2000 offers up to 400 FP4 TFLOPS with an Nvidia Blackwell GPU featuring 1,024 CUDA cores and 16 GB of LPDDR5X memory.
What applications is the T2000 aimed at?
The T2000 is aimed at applications including visual AI agents, autonomous mobile robots and robotic manipulators.

Get the latest Robotics news every morning

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Free · takes 30 seconds · unsubscribe anytime

Discussion

No comments yet. Be the first to share your thoughts!

Log in to join the discussion

Related Articles

Stay ahead with AI news

Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.

Get Started Free

Free · takes 30 seconds · unsubscribe anytime

1 minute a day. The AI essentials.

200+ sources · Email / LINE / Slack

Get it free →