
Applications
AI(AIoT)
AI meets IoT in wearables and smart environments — and ASE's advanced packaging brings compute, connectivity and sensing into an ultra-light frame.
Overview
AI and IoT are converging in wearables and smart environments everywhere.
In an era where connectivity and intelligence are paramount, the convergence of artificial intelligence (AI) and the Internet of Things (IoT) has given rise to a transformative phenomenon known as the Artificial Intelligence of Things (AIoT). This fusion is reshaping not only wearable technology, but also entire ecosystems where data-driven intelligence is embedded into everyday environments. Wearables such as AI smart glasses, smartwatches, and smart rings are evolving from basic fitness and communication tools into sophisticated platforms that enhance health monitoring, productivity, and daily interactions—while extending intelligence to broader use cases like Smart City infrastructure, connected public safety systems, and responsive Smart Home automation.
As AIoT expands across these domains, several key trends and innovations are emerging. In semiconductor packaging, high-density heterogeneous integration (HI) is enabling more capable, power-efficient edge inference by tightly combining compute, memory, sensing, and RF functions. At the system level, enhanced connectivity through 5G networks supports low-latency interaction and richer real-time data sharing, which is essential for Smart City and industrial applications as well as advanced wearable experiences. Meanwhile, the development of comfortable physiological sensing technologies—integrating improved signal quality, miniaturized sensors, and intelligent interpretation—helps make continuous health and context-aware monitoring practical for Smart Wearables.
Despite these advances, AIoT deployment faces significant challenges, particularly in testing methodologies, radio frequency (RF) co-verification, and thermal management. High integration density and mixed-signal complexity make it harder to validate performance consistently across devices and production lots, while maintaining reliable RF behavior becomes more critical as connectivity demands increase. At the same time, thermal constraints directly affect comfort, safety, and long-term reliability for wearables and compact edge nodes. Addressing these issues is crucial to ensure AIoT products—spanning Smart Cities, Smart Homes, and Smart Wearables—deliver the performance, usability, and comfort expectations users rely on.
trends & innovations
Denser integration, 5G and comfortable sensing are driving the next wearables.
The most important part of AIoTs is about wearable products, especially AI smart glasses, smartwatches, and smart rings. These devices are moving from simple tools for fitness or communication to more advanced products that improve health monitoring, work efficiency, and daily interaction. AI glasses are now used not only for gaming or entertainment but also for training, design, and remote teamwork, thanks to the integration of AI, allowing for real-time data analysis, enabling users to make informed health decisions and enhancing overall well-being.
High-density heterogeneous integration
In the last few years, high-density heterogeneous integration (HI) packaging has become a core trend in advanced semiconductors. This refers to combining multiple chips (like processors, memory, power management, sensors) into extremely compact modules using approaches like Fan-Out Wafer Level Packaging (FOWLP), 2.5D and 3D chip stacking, and System-in-Package (SiP) designs.
These methods help shrink size and improve power performance because signals travel shorter distances and heat can be controlled better. Such integration is key for wearable products, especially AI glasses, which need to squeeze high compute, graphics, and sensors into a tiny frame while managing heat and power effectively.
Enhanced connectivity with 5G
The rollout of 5G networks is a game-changer for AIoT applications. With significantly higher data transfer speeds and lower latency, 5G enables seamless connectivity between devices, facilitating the exchange of large volumes of data in real-time. This enhanced connectivity is crucial for applications such as AI glasses, to support independent, connected AI glasses that work without a phone, deliver low latency and high speed data, and maintain comfortable form factors while meeting wireless standards.
High-frequency RF (radio frequency) packaging is especially important for AI glasses with standalone 5G connectivity. To support communications directly on the device, the package must not only house RF chips but allow precise test verification of performance across the entire frequency range, including impedance matching, transmission loss and electromagnetic compatibility.
Comfortable physiological sensing
The next generation of wearable devices is set to incorporate advanced biosensors, such as those for heart rate, eye tracking, blood oxygen levels, and other biological signals. However, a key challenge lies in integrating these sophisticated biometric sensors into wearable technology without sacrificing user comfort, freedom of movement, or the overall aesthetics of the device.
Given the limited contact surface area of wearable devices, it is crucial to tightly package and rigorously test the electronics to ensure they function accurately without hindering motion or causing discomfort. Additionally, it is important to minimize interference with optical components. To achieve this, a flexible optical sensor that can adapt to the contours of the wearer's face will be essential. Such a sensor can maintain stable contact with the skin, enhancing the reliability of physiological data collection, all while remaining compact and lightweight.
Key issues
Testing, RF co-verification and heat are the challenges to solve.
Advanced test methodologies
The economic feasibility of high-density heterogeneous integration for advanced wearables hinges on the sophistication of the backend testing infrastructure. As devices integrate multi-die logic, power management, high-frequency RF, and environmental sensors into a single ultra-miniaturized package, traditional legacy testing methods become inadequate.
To avoid catastrophic yield loss—where a single defective 40nm analog filter can render a pristine 4nm application processor within a System-in-Package (SiP) unusable—advanced testing capabilities are crucial. These include high-speed signal integrity testing, thermal behavior characterization, and built-in self-test (BIST) to ensure consistent performance in high-volume production. Additionally, the integration of simulation models with empirical test data facilitates precise localization and root-cause analysis of internal defects.
Radio frequency (RF) co-verification
Modern wearables require simultaneous parametric testing across a variety of mixed-signal domains. These wearable devices are densely packed with wireless communication standards, including 5G, Bluetooth Low Energy (BLE), Wi-Fi 7, Ultra-Wideband (UWB), and Near Field Communication (NFC). Testing these radios within a highly compact package presents significant challenges due to electromagnetic interference (EMI) from adjacent digital switching circuits. High-frequency signals are particularly sensitive to the design, materials, and geometry of the package. Traditional low-frequency tests are ineffective, while high-frequency or mmWave testing is costly, complex, and influenced by environmental factors, complicating the assurance of real-world 5G performance.
To address these challenges, OSATs must implement multi-site RF testing using shielded, isolated test chambers integrated into the device handler. Additionally, setups involving Vector Network Analyzers (VNA) are essential for measuring high-frequency performance, such as S-parameters and impedance matching. System-level RF testing should be conducted to evaluate the RF module, antenna, and package collectively, providing a more accurate simulation of real-world performance. Furthermore, combining simulation with measurement—utilizing RF simulation tools and comparing the results with actual measurements—can enhance test accuracy.
Thermal management in "touch-sensitive" AIoT devices
AIoT devices are shrinking even as their thermal envelopes expand. High-performance wearables, in particular, generate considerable heat during peak AI inference, making it challenging to dissipate this energy effectively within 3D-stacked architectures. This leads to "thermal throttling," where the AI performance is capped to prevent hardware damage. Moreover, while a smartphone back cover can safely hit 45°C, wearables like AI glasses sit against the temple—a zone uniquely sensitive to heat.
To address this issue, "Thermal Bridges" can be designed and embedded within the System-in-Package (SiP) to redirect heat away from the skin and toward the front of the frame. Additionally, the industry is increasingly adopting Digital Twin-based Thermal Modeling. By simulating the thermal behavior of the package before physical production, engineers can optimize substrate materials and heat sink placement with precision.
ASE technology platform
VIPack, SiP and WLP pack a full system into an ultra-light frame.
VIPack™ — Integrated Design Ecosystem
Backed by a comprehensive, integrated design ecosystem (IDE), VIPack™ allows engineers to simulate, plan, and optimize the package during the earliest stages of development. IDE integrates multiphysics simulations (electrical, thermal, warpage/stress, reliability) with AI-based risk prediction and real-time manufacturing data correlation, enabling faster design-analysis cycles and data-driven optimization. When an internal defect is suspected, correlated simulation-to-test datasets allow failure analysis engineers to pinpoint the root cause with minimal trial-and-error, reducing yield-learning cycles and keeping high-volume production on track.
SiP / WLP
Our SiP (System in Package) and WLP (Wafer Level Packaging) can be used to implement compact RF front-end modules and integrate antennas, filters, and power amplifiers together to reduce size and power for mobile and wearable devices. These capabilities are essential for space-constrained AIoT devices—smart AI glasses, for instance—that must pack high-performance computing, advanced graphics, and sensor arrays into an ultra-light frame while managing strict thermal and power limits.
VIPack™ — 2.5D/3D IC & FOCoS/FOSiP
VIPack™ is ASE's advanced packaging platform designed to implement high-density heterogeneous integration. It combines technologies like high-density RDL (redistribution layers), advanced fan-out packaging (FOCoS/FOSiP), and TSVs (through-silicon vias) of 2.5D/3D ICs. This allows multiple chips to be tightly stacked, delivering maximum interconnect density, ultra-low power consumption, and peak performance in a minimal footprint. Such integration is key for space-constrained AIoT devices, such as smart AI glasses, which demand high-performance computing, advanced graphics, and diverse sensor arrays packed into an ultra-light frame, all while managing strict thermal and power constraints.
Conclusion
Advanced packaging and testing will define the future of AIoT.
The Artificial Intelligence of Things represents a significant leap forward in wearable technology, merging the capabilities of AI with the connectivity of IoT to create devices that are not only smarter but also more integral to our daily lives. As we witness the rapid evolution of AIoT, the focus on high-density heterogeneous integration, enhanced connectivity through 5G, and the incorporation of advanced biosensors will be pivotal in shaping the future of wearables. However, the challenges associated with testing, RF performance, and thermal management must be addressed to ensure these devices operate reliably and efficiently.
By leveraging advanced packaging technologies like ASE's VIPack™ and adopting innovative testing methodologies, the industry can overcome these hurdles, paving the way for a new era of AIoT devices that enhance user experience and promote well-being. As we move forward, the potential of AIoT to revolutionize health monitoring, productivity, and personal interaction is immense, promising a future where technology seamlessly integrates into our lives, enhancing our capabilities and enriching our experiences.




