NVIDIA Jetson Thor Developer Kit Blackwell - Supercomputer for AI and Robotics
NVIDIA Jetson Thor Developer Kit Blackwell - Supercomputer for AI and Robotics
Couldn't load pickup availability
ROI Kalkulačka
Spočítejte si návratnost investice.
SKU: 1601549049087
Description
- ✔ Extreme AI performance: Up to 2070 TFLOPS (FP4—Sparse)
- ✔ Revolutionary GPU: NVIDIA Blackwell architecture with 2560 cores
- ✔ Robust CPU: 14-core Arm® Neoverse®-V3AE 64-bit processor
- ✔ Huge memory: 128GB LPDDR5X for complex tasks
- ✔ Lightning-fast storage: 1TB NVMe M.2 Key M for instant data access
- ✔ Top connectivity: Up to 32 cameras and rich I/O options
✨ The innovative heart of your AI projects: Key features ✨
- Blackwell Architecture: NVIDIA's latest GPU with 2560 cores and 96 fifth-generation Tensor cores, delivering massive AI acceleration.
- Powerful 14-core CPU: Arm® Neoverse®-V3AE 64-bit processor with 1 MB L2 cache per core and 16 MB shared L3 cache to smoothly run even the most demanding tasks.
- Extensive memory and storage: 128GB LPDDR5X memory and a lightning-fast 1TB NVMe M.2 Key M slot for plenty of space and speed for data and models.
- Cameras and Sensors: Flexible support for up to 32 cameras with MIPI CSI-2 and C-PHY 2.1 for sophisticated ambient perception, crucial for robotics and autonomous systems.
- Modern connectivity: Interfaces such as 5GBe, QSFP28 (4x 25 GbE), PCIe Gen5 slots (M.2 Key M and Key E), and a wide range of I/O for maximum flexibility.
🔍 Explore every detail – Uncovering your full potential 🔍
The NVIDIA Jetson Thor Developer Kit with Blackwell architecture represents a major milestone in the field of edge AI and robotics. This platform is designed for developers, researchers, and engineers looking for uncompromising performance for the most demanding AI tasks right on the device. With a Blackwell GPU that achieves 2070 TFLOPS in FP4 sparse precision, the kit is capable of processing massive amounts of data in real time, which is essential for complex applications such as autonomous vehicles, humanoid robots, and intelligent industrial systems.
The integrated 14-core Arm® Neoverse®-V3AE processor complements the graphics performance and ensures smooth operation of the operating system and parallel computing tasks. With 128 GB LPDDR5X memory and 1 TB NVMe storage, the kit has sufficient reserves to work with large datasets and complex AI models. A wide range of input and output interfaces, including support for up to 32 cameras, multiple PCIe Gen5 slots and high-speed Ethernet ports, allows for the connection of a wide range of sensors, actuators and peripherals.
Rugged construction with a Thermal Transfer Plate and optional fan or heatsink ensures reliable operation even under full load. The Jetson Thor Developer Kit is the ultimate tool for those looking to push the boundaries of artificial intelligence and robotics.
🤔 Are you worried about limited computing power on Edge devices? 🤔
Do you feel like your existing embedded systems can't keep up with demanding AI models and real-time data processing applications? Do you need a platform that offers data center-level performance in a compact form factor for field deployment? Limited connectivity and memory often hinder the full development of your innovative robotics and autonomous systems projects.
🚀 Gain the upper hand with the NVIDIA Jetson Thor Developer Kit! 🚀
The NVIDIA Jetson Thor Developer Kit is designed to address these challenges. It delivers unmatched AI performance with the Blackwell architecture, which enables you to run even the most complex neural networks with minimal latency. With this kit, you gain the freedom to innovate and deploy cutting-edge AI applications in the real world, whether it’s advanced robotics, autonomous systems, or industrial automation, where speed and accuracy are key.
🌟 More than just features: Real benefits you'll feel 🌟
- Extreme AI acceleration: 2070 TFLOPS (FP4) -> Benefit: Perform even the most complex AI inference and data analysis in real time directly on the device.
- Compact data center power: Blackwell GPU and 14-core CPU -> Benefit: You get top-notch computing power in a small form factor, ideal for embedded and Edge AI applications.
- Flexible environmental perception: Support for up to 32 cameras -> Benefit: Develop sophisticated autonomous systems with complex visual perception.
- Future-ready: PCIe Gen5, 5GBe, QSFP28 -> Benefit: Ensures compatibility with the latest peripherals and ultra-fast communication for your projects.
- Extensive expansion options: M.2 slots, rich I/O -> Benefit: Easily integrate additional sensors, communication modules and storage according to your specific needs.
🎯 Who is the NVIDIA Jetson Thor Developer Kit an ideal choice for? 🎯
- AI and machine learning developers: For those who need a robust platform to deploy complex models at the Edge.
- Robotics Engineers: Ideal for developing advanced humanoid robots, autonomous mobile robots (AMRs), and industrial manipulators.
- Scientists and researchers: Perfect for research in the areas of computer vision, autonomous driving, and advanced artificial intelligence.
- Companies innovating in IoT and Industry 4.0: For creating intelligent production lines, monitoring systems, and smart devices.
💬 Do you have concerns? We'll answer them! 💬
- Concern: Is such a powerful platform difficult to program? -> Answer: NVIDIA provides an extensive software stack, including JetPack SDK, CUDA, and TensorRT, that simplify the development of AI applications and their optimization.
- Concern: How is cooling of such power ensured? -> Answer: The kit is equipped with an integrated Thermal Transfer Plate (TTP) with heat pipes and optional fan or heat sink support for effective heat dissipation.
- Concern: Is the platform ready for future technologies? -> Answer: With PCIe Gen5, 5GBe, and support for the latest standards, Jetson Thor is designed with long-term sustainability and compatibility with future innovations in mind.
⚙️ Technical specifications and package contents ⚙️
| AI performance (FP4 Sparse) | 2070 TFLOPS |
| GPU architecture | NVIDIA Blackwell |
| GPU cores | 2560-core NVIDIA Blackwell architecture GPU with 96 fifth-generation Tensor Cores |
| Multi-Instance GPU (MIG) | Yes, with 10 TPC |
| CPU | 14-core Arm® Neoverse®-V3AE 64-bit CPU |
| CPU L2 cache | 1 MB per core |
| L3 system cache | 16 MB shared |
| Memory | 128GB 256-bit LPDDR5X |
| Memory throughput | 273 GB/s |
| Internal storage | 1TB NVMe M.2 Key M Slot |
| NVMe support | Via PCIe |
| SSD support | Via USB3.2 |
| HSB camera input | Via QSFP slot |
| Total number of HSB cameras | Up to 20 |
| MIPI CSI-2 cameras | Up to 6 over 16x lanes |
| Virtual camera channels | Up to 32 cameras with C-PHY 2.1 (10.25 Gbps) |
| M.2 Key M slot | x4 PCIe Gen5 |
| M.2 Key E slot | x1 PCIe Gen5 (WLAN/BT, x1 PCIe, USB2.0, UART, I2C, I2S) |
| PCIe Gen5 | Up to x8 lanes (Root port only—C1 (x1) and C3 (x2)), Root Point or Endpoint—C2 (x1), C4 (x8), and C5 (x4) |
| USB host controller | xHCI with integrated PHY |
| Ethernet | 1x 5GBe RJ45 connector, 1x QSFP28 (4x 25 GbE) |
| Video output | VESA DisplayPort 1.4a—HBR2, MST |
| CAN header | 2x 13-pin |
| Automation header | 2x 6-pin |
| JTAG connector | 2x 5-pin headers |
| Fan connector | 1x (12V, PWM, Tach) |
| Audio panel header | 2x 5-pin |
| RTC battery connector | 2-pin |
| Low-speed I/O | 5x I2S/2x audio hub (AHUB), 2x DMIS, 4x UART, 4x CAN, 3x SPI, 13x I2C, 6x PWM outputs |
| Dimensions | 243.19mm x 112.40mm x 56.88mm |
| Cooling | Integrated Thermal Transfer Plate (TTP) with heatpipe, optional fan or heatsink |
| B2B connector | 699-pin |
❓ Frequently Asked Questions (FAQ) ❓
- What applications is Jetson Thor ideal for?
- It is ideal for autonomous robotics, humanoid robots, AI-accelerated Edge devices, intelligent industrial automation, and advanced analytics systems that require extreme performance.
- What is the main difference between Jetson Thor and previous Jetson platforms?
- Jetson Thor delivers the new Blackwell architecture with significantly higher AI performance (up to 2070 TFLOPS FP4 sparse), a more powerful 14-core CPU, and expanded connectivity and memory options (128 GB LPDDR5X) compared to previous generations.
- What operating systems and software stack are supported?
- The platform supports the NVIDIA Jetson Linux Developer Guide, which includes the JetPack SDK, CUDA, TensorRT, and other AI tools for easy application development and deployment.
- Can I connect external storage to the Developer Kit?
- Yes, the kit features a 1TB NVMe M.2 Key M slot for internal storage and supports NVMe over PCIe as well as SSD over USB3.2, providing ample expansion options.
- How many cameras can I connect to Jetson Thor?
- Jetson Thor supports up to 20 HSB cameras via QSFP slot and up to 6 cameras via 16x lanes MIPI CSI-2. Up to 32 cameras can be connected using virtual channels with C-PHY 2.1.
- What are the network connection options?
- The kit offers 1x 5GBe RJ45 connector and 1x QSFP28 (4x 25 GbE), which ensures ultra-fast network connectivity for transferring large data streams.
- Is it possible to connect wireless modules?
- Yes, the M.2 Key E expansion slot with x1 PCIe Gen5 supports WLAN/Bluetooth modules, as well as USB2.0, UART, I2C and I2S interfaces for expanding wireless connectivity.