At this week’s Embedded World in Anaheim, Quectel launched three smart modules built on MediaTek’s Genio platform.

The SH603FC, SH503FM and SH803FD bring octa-core processing and dedicated on-device AI acceleration to a broad range of connected products, without the cost or complexity of integrated cellular connectivity.

Together, the three modules form a family spanning entry, mid and high-performance tiers, giving manufacturers an upgrade path as their compute and AI requirements grow. Each is available in a wifi variant, for products that need built-in wireless connectivity, and the SH603FC is available in a pure application processor variant, for designers pairing the module with their own connectivity.

All three share a common architecture: an octa-core Arm Cortex-A78 and Cortex-A55 processor, Arm Mali-G57 MC2 GPU, and support for Android 15 and Yocto Linux (Kernel 6.6). Each integrates LPDDR5X memory up to 16Gbyte and dual independent display output via Mipi (LVDS), eDP or DP.

All three are designed without integrated cellular connectivity, giving users flexibility to choose connectivity that best fits their needs while keeping costs and certification requirements streamlined. All three are qualified to commercial and industrial grade, with long-term supply commitments built in.

The combination of on-device AI, rich multimedia support and dense interfaces – including LCM, camera, touchscreen, microphone, uart, USB, I²C, SPI, RGMII and PCIe – suits products such as industrial smart tablets, smart intercom and recorder devices, edge computing equipment, vending machines, logistics lockers, smart PoS systems, aftermarket in-vehicle entertainment, digital signage, companion smart displays, edge AI gateways, camera boxes, robotics, and smart home systems.

“These three modules give our customers a single, trusted platform to scale from straightforward connected displays through to demanding edge AI applications,” said Raymond Wang from Quectel (www.quectel.com). “Whether a customer needs a cost-effective panel for a vending machine or the processing power for on-device video analytics, the Genio-based family means they don’t have to redesign around a new architecture every time their requirements change.”

The SH503FM (www.quectel.com/product/wi-fi-bt-sh503fm-smart-module-series/) is the family’s mid-tier module, based on the Genio G420 (MT8371LV) platform. Its NPU delivers up to 7.2Tops, pairing a single-ISP camera pipeline with the same 16MP at 30fps capture and 4K at 60fps playback as its siblings, making it a natural fit for single-camera edge AI use cases.

Offered as the SH503FM-WF or SH503FM-AP, the SH503FM targets PoS systems, handheld terminals, smart home panels and human-machine interfaces where light edge inference, such as voice or basic object detection, adds value without the cost of a higher-tier processor.

The SH603FC (www.quectel.com/product/sh603fc-wi-fi-bluetooth-smart-module-series/) is the mid-to-high-end entry point in the family, built on the Genio G520 (MT8371) platform with an NPU delivering up to 10.3Tops of on device AI compute. The module is available with Wifi 6E and Bluetooth 5.4 or Wifi 6 and Bluetooth 5.0 and supports up to 16MP camera input at 30fps and 4K video recording and playback at 60fps, alongside five USB 3.0/2.0, eight I²C, five SPI and RGMII interfaces.

It is suitable for industrial smart tablets, edge computing boxes, smart cash registers and digital signage that need strong multimedia performance without the premium price of a top-tier module.

The SH803FD (www.quectel.com/product/wi-fi-bt-sh803fd-smart-module-series/) sits at the top of the family, built on the Genio G720 (MT8391) platform with an NPU delivering up to 10.5Tops, the highest in the range. Available with Wifi 6E and Bluetooth 5.4, it adds a dual-ISP camera pipeline supporting 16MP plus 16MP capture at 30fps, 4K at 60fps video decoding, 4K at 30fps encoding, and PCIe 2.0, USB 3.2 and SD 3.0 interfaces.

The SH803FD is designed for demanding applications such as multi-camera edge computing, industrial rugged tablets, robotics and generative AI workloads, including on-device meeting summarisation and lightweight local language models.