IoT compute specialist Lantronix has launched five new System-in-Package (SiP) solutions designed to bring AI and Machine Learning capabilities to the edge. The new SiP packages are powered by Qualcomm Technologies’ chipsets and result from close collaboration between the two companies that has extended over fifteen years.

The five new SiP families are designed to accelerate the development of AI-driven applications in industrial and enterprise use cases, including robotics, industrial automation, video surveillance, video collaboration and drones. The new SiP modules are compliant with the US Trade Agreements Act (TAA) and the National Defense Authorization Act (NDAA).

Lantronix has integrated AI capabilities from the Qualcomm AI Hub which provides a reference base of more than 100 AI models and a simplified model optimisation process. This efficiently uses AI capabilities (3.5 to 100 INT-8 TOPS) in the new Lantronix SiP families.  

Lantronix’s pin-compatible 9100IQ and 9075IQ SiPs, powered by the Qualcomm IQ-9100 and IQ-9075 processors, are designed for scalable, power-efficient computing for autonomous devices and Industry 4.0 designs using AI. The new IQ9 Series can enable:

• Robust safety functions in autonomous mobile robots (AMR) or platforms with functional safety (FuSa) up to level SIL-3 level (IQ-9100-based SiPs only)
• Device robustness with fault tolerance Error Correction Code (ECC) memory support and system cost savings through an integrated, dedicated safety island (IQ-9100) or real-time subsystem (IQ-9075) with four dedicated independent processing cores supporting real-time operating systems for system error monitoring and other critical functions.
• Robot perception, navigation and versatility improvement through a Qualcomm Adreno 663 GPU and support for up to 16 concurrent cameras.
• Interactive industrial edge AI systems using up to 100 TOPS by integrating Large Language Model (LLM) support at the edge.
  The IQ9 Series Hexagon tensor processor can achieve a generation rate of 12 tokens per second when running the Llama 2 13B parameter mode.
• Fanless systems to enhance operating temperature with the SiP family supporting a -40°C to 115°C junction temperature range.

The Lantronix’s Open-Q 8550CS family, powered by Qualcomm Technologies’ QSC8550 processor, is designed for high AI performance, power efficiency and Wi-Fi 7 and Bluetooth 5 connectivity. Benefits include the abilities to:

• Enhance video conferencing meeting experiences, automated guided vehicle pathing, smart camera image quality and edge AI box scalability with the family’s octal-core computing capabilities and 48 AI TOPS tensor performance.
• Perform complex 3D rendering and computer vision tasks with a Adreno 740 GPU supporting ray tracing, Open GL ES, Vulkan and Open CL profiles and 4K240/8K60 video decoding and 4K120/8K30 encoding.
• Connect edge AI boxes leveraging high-speed 2.5G and 10G Ethernet ports.

Lantronix’s pin-compatible Open-Q 6490CS and Open-Q 5430CS families, powered by Qualcomm Technologies’ QCS6490 and QCS5430 processors, is designed to allow users to scale product lines with low development effort and receive low-power AI performance, Wi-Fi 6E and BLE 5+ connectivity. Features include:

• Real-time machine learning on 6th-generation AI engine, delivering 3.5 to 13 AI TOPS and complemented with up to octal-core CPU and Adreno 640 class GPU. 
• Advanced multimedia and AI powered camera support through up to three concurrent ISPs supporting up to 192MP cameras, 4K30 encoding and 4K60 decoding, sufficient to handle up to 8 camera streams simultaneously for video-intensive applications.
• Percepxion device management for over-the-air (OTA) upgrades for performance, security and software feature improvements. 

“Qualcomm Technologies and Lantronix have had strong relationships for more than 15 years,” said Dev Singh, vice president of Business Development and head of Industrial Automation at Qualcomm Technologies Inc.
“Utilising Qualcomm Technologies’ cutting-edge processors, Lantronix enables its customers to seamlessly deploy AI solutions at the edge, bringing its expertise in embedded computing and IoT to deliver reliable, industrial-grade systems.”

“With the addition of these five new SiP solutions, we continue our strategic collaboration with Qualcomm Technologies that has enabled Lantronix to build a proven track record of successfully delivering integrated, collaborative solutions that are driving forward IoT and AI/ML technologies to meet the evolving needs of today’s advanced-edge applications,” said Mathi Gurusamy, chief strategy officer for Lantronix.