Exhibitor: Ambiq Micro
Hall/Booth: 4-581
Ambiq HeliaAOT is an ahead of time compiler solution that enables modern AI models to run efficiently on highly resource constrained edge devices. While traditional AI runtimes introduce unnecessary overhead, high memory demands, and unpredictable memory usage, HeliaAOT converts trained models—such as TensorFlow Lite—into highly optimized, standalone C inference modules already at build time.
By eliminating a generic runtime, removing unused operators, and providing automatic, deterministic memory planning, HeliaAOT produces extremely compact, transparent, and energy efficient deployments. Hardware optimized transformations such as layer fusion, tensor reordering, and graph based optimizations significantly reduce code size and memory footprint while still enabling powerful, real time inference performance.
Developed specifically for Ambiq’s ultra low power SoCs with SPOT® technology, HeliaAOT delivers up to 10× smaller binaries, predictable execution, and industry leading efficiency—ideal for wearables, hearables, sensors, and other always on devices. With reduced energy consumption, smaller memory requirements, and extended device lifecycles, HeliaAOT also contributes to sustainable, resource efficient edge AI systems.
Exhibitor: Ceva
Hall/Booth: 4-462
Ceva NeuPro Nano is a fully self sufficient Neural Processing Unit (NPU) that brings powerful, energy efficient AI directly to highly resource constrained edge devices. While traditional solutions depend on additional host CPUs or DSPs—adding complexity, power consumption, and cost—NeuPro Nano integrates neural inference, control logic, signal processing, and memory management into a single programmable NPU core.
Supporting modern ML data formats from 4 to 32 bit, native transformer processing, hardware accelerated sparsity, and fast quantization, NeuPro Nano delivers scalable performance with exceptionally low energy use. The proprietary Ceva NetSqueeze technology enables direct processing of compressed weights, reducing model memory requirements by up to 80 %.
With ultra low power operation, deterministic latency, and a compact system architecture, NeuPro Nano is ideal for always on applications in consumer IoT devices, industrial sensing, and automotive systems. Its efficiency extends battery life, reduces material needs, and minimizes cloud dependency through local AI inference.
NeuPro Nano offers a future proof, sustainable, and cost efficient platform for physical AI applications right at the sensor—delivering maximum performance with minimal resources.
Exhibitor: 7Starlake Co., Ltd.
Hall/Booth: 1-352
The 7Starlake Counter UAV System delivers modern drone protection for military and other high security environments. Combining rugged, mission ready hardware with AI powered analytics, it reliably detects, tracks, and assesses unauthorized or hostile UAVs. The AV600 RH A45 Military GPU Computer and the Cloud15 PX6 Panel PC provide high performance processing even under extreme conditions.
Using AI acceleration through the OpenVINO™ toolkit, the system analyzes radar, optical, and thermal sensor data in real time, offering precise situational awareness and automated decision support. Integrated sensor fusion enhances accuracy and response speed, enabling early threat recognition and rapid action.
Its durable, modular hardware architecture is designed for sustainable long term deployment. Energy efficient components, robust construction, and upgradeable modules reduce maintenance needs and overall resource consumption.
The result: an intelligent, resilient, and future ready defense system that reliably protects critical infrastructure from emerging UAV threats.
