FalconS1 Pro is an edge AI computing platform designed to make high-performance inference practical, scalable, and easy to deploy in real-world industrial environments.

At its core, FalconS1 Pro integrates the NXP i.MX8M Plus processor with a Hailo AI accelerator delivering up to 26 TOPS of AI performance. This combination enables efficient, low-latency inference directly at the edge, removing the need for constant cloud dependency.

Unlike general-purpose embedded boards, FalconS1 Pro is built specifically for industrial and production-grade AI workloads. It combines compute power with a full suite of industrial interfaces and a software stack designed to simplify AI deployment through our AI Studio and AI-as-a-Service model deployment system.

Keep Your Insights Local

Modern edge computing systems increasingly require real-time decision-making close to where data is generated. FalconS1 Pro is designed for these environments, enabling developers and system integrators to process sensor data, run analytics, and deploy ML models directly at the edge. In manufacturing, it can perform real-time quality inspection and defect detection on production lines. By analyzing data locally, manufacturers can reduce latency, improve reliability, and maintain operation even when cloud connectivity is limited.

The platform is also well suited for predictive maintenance applications. Data from vibration, temperature, current, pressure, and acoustic sensors can be processed directly on the device to monitor equipment health and detect anomalies before failures occur. This helps reduce downtime and enables more efficient maintenance scheduling.

Beyond manufacturing, FalconS1 Pro can be used for environmental and infrastructure monitoring. The board can collect and analyze data from air-quality sensors, gas sensors, weather stations, energy meters, and other industrial sensors to provide real-time insights and automated responses. Combined with AI acceleration, it can support advanced workloads such as anomaly detection, forecasting, and sensor fusion.

To further simplify deployment, developers can optionally use our EdgeMind AI Studio to create and manage AI workflows, AI Model-as-a-Service (MaaS) to organize and deploy optimized models to accelerate common industrial use cases. These tools complement the hardware while keeping inference running locally on FalconS1 Pro.

Features & Specifications

Core Hardware

• NXP i.MX8M Plus processor
• Hailo AI accelerator (up to 26 TOPS)
• 8 GB RAM / 128 GB storage configuration
• Industrial-grade edge AI design

Software Stack

• Yocto-based/Ubuntu Linux OS
• Docker support for containerized deployment
• EdgeMind AI Studio for low-code/no-code AI workflow development (cloud or on-premise deployment)
• AI Model-as-a-Service (MaaS) for AI model management and deployment
• Runtime applications including OCR, with additional industrial applications such as Energy Management System (EMS) and Overall Equipment Effectiveness (OEE) under development
• AI inference runs locally on FalconS1 Pro for low-latency operation

Connectivity & Industrial Interfaces

• Full industrial peripheral support (UART, SPI, I2C, GPIO, CAN, etc.)
• Designed for integration into embedded and factory systems
• Edge-first architecture for low-latency workloads

General

• Optimized for continuous industrial operation
• Designed for scalable deployment across multiple edge nodes
• Optional software ecosystem to simplify AI development, deployment, and lifecycle management

Open Source

FalconS1 Pro follows an open development approach to enable transparency and community-driven innovation.

• Hardware schematics are open source
• Yocto BSP is open source
• Software SDK components are publicly available
• Project repositories are hosted on a public GitHub organization

Developers and partners can access design files, board support packages, firmware, and SDK resources to build custom solutions or contribute improvements. FalconS1 Pro also integrates with our broader software ecosystem, allowing developers to choose the workflow that best fits their application while maintaining local AI inference on the device.

Open-source resources will be available prior to production and will continue to be updated throughout development and deployment phases.

Interested in This Project?

You can sign up at the top of this page to be notified when the campaign launches and to receive other updates. We only send relevant content, and you can unsubscribe at any time.