Taipei, Taiwan – October 10, 2025 – AVerMedia Technologies, Inc., a leading global provider of application-ready Edge AI turnkey solutions, today announced the launch of the AI Fusion Kit, an application-ready multimodal input solution for developers building language and vision-language models (LLMs & VLMs) at the edge.
Product Overview
The AI Fusion Kit bundles AVerMedia's Jetson-based Box PC, a 4K high-resolution USB camera, a high signal-to-noise ratio (SNR) speakerphone, pre-installed storage, and a Jetson-optimized software stack into a single turnkey kit. The solution streams synchronized, high-fidelity audio and video directly into local AI inference pipelines, enabling more accurate speech and vision inputs for multimodal conversational agents and VLM applications.
"The AI Fusion Kit is designed to remove friction from edge AI development," said Alex Liu, Vice President of Industrial Product Division at AVerMedia. "By combining validated hardware, synchronized multimodal capture, and a ready-to-run software stack, we help developers accelerate prototypes into pilots — reducing integration effort, shortening time-to-insight, and enabling enterprises to deliver more capable AI at the edge."
Product Family
The AI Fusion Kit is available in two SKUs to match different performance and power requirements:
- Fusion Kit 1 — AGX Orin edition (NVIDIA Jetson AGX Orin, 32 GB)
Designed for the highest on-device compute for demanding LLM/VLM workloads. - Fusion Kit 2 — Orin NX edition (NVIDIA Jetson Orin NX, 16 GB)
Optimized for compact, power-efficient multimodal deployments.
From top to bottom: AVerMedia AGX Orin edition Box PC, Orin NX edition Box PC, high-SNR speakerphone, and 4K USB webcam.
Key Features and Benefits
- Ready-to-run multimodal inputs (4K camera options; high-SNR speakerphone) for robust speech and vision capture.
- Jetson-preactivated Box PC with a BSP-optimized software stack (JetPack-based) to speed initial setup.
- Containerized GUI demo applications, a quick-start guide, and setup scripts to accelerate onboarding.
- Broad model support for common LLM/VLMs (Llama, Phi-4, LLaVA, etc.) to simplify model integration.
- Validated reference integrations to accelerate PoCs and reduce engineering effort for pilots and deployments.
- Flexible I/O and expansion headers for connecting networking and cellular modules, and multiple camera input support (depending on SKU).
Application
Designed for real-world edge development and pilot operations, the AI Fusion Kit is particularly well-suited for smart retail and security projects that require synchronized, high-quality audio/video capture, as well as reliable local inference.
Availability and Resources
The AI Fusion Kit (both editions) will be available beginning October 10, 2025. For sales inquiries, evaluation kits, or PoC support, contact AVerMedia. To learn more about the AI Fusion Kit, visit:
- Fusion Kit 1 (AGX Orin edition):
https://professional.avermedia.com/product-detail/D315AOB-2-Fusion-kit - Fusion Kit 2 (Orin NX edition):
https://professional.avermedia.com/product-detail/D133SOXB-Fusion-kit - Quick Start Guide:
https://developer.avermedia.com/blog/ai-fusion-kit-quick-start-guide/ - Unboxing video:
https://www.youtube.com/watch?v=Uwdl8Nzj25w
About AVerMedia
AVerMedia Technologies, Inc. is a leading global provider of application-ready Edge AI turnkey solutions, recognized for its expertise in video and audio technologies. AVerMedia offers customized, fully integrated hardware and software solutions designed to accelerate AI deployments across various industries, including smart cities, robotics, and industrial automation. As an NVIDIA Elite Partner, AVerMedia is dedicated to delivering cutting-edge AI inference solutions that enable businesses to scale AI effectively and reliably at the edge.
Learn More: https://professional.avermedia.com
Follow Us: https://www.linkedin.com/showcase/avermedia-edge-ai/


