Jetson is an embedded system platform developed by NVIDIA to run AI applications directly on the device (edge AI). By processing data locally on the device, robots, drones, or smart cameras can respond in real time, operate without relying on an internet connection, and securely handle data without sending it outside. For example, an autonomous vehicle that detects a sudden obstacle can instantly process the data with Jetson and brake or change direction with low latency, without sending the information to a remote server over the internet.

The Jetson platform consists of two main product groups designed for different use cases:

1. Modules: Embedded processing units that include GPU, CPU, memory, and necessary interfaces. These modules define the computing power of AI applications and are designed for long-term, reliable use in production environments. To become functional, they must be integrated onto a compatible carrier board.

2. Developer Kits: Ready-to-use sets that include a Jetson module, a reference carrier board, and basic I/O connections. They are designed for AI project development, testing, and prototyping, offering easy setup. They are intended more for R&D purposes rather than long-term production.

For users who want to deploy Jetson in long-term production environments, there are two options:

• Integrating components yourself: Jetson modules, compatible carrier boards, and other required hardware can be assembled to create custom systems. This requires technical expertise.
• Choosing ready-made solutions: Pre-integrated AI kits or boxed systems can be used. This speeds up the setup process and allows for an easier transition to production environments.