Hyperscaler Capex keeps rising on price hikes and shortages; pricier AI racks drive server value growth over volume.
US hyperscalers keep expanding, but midterm-driven backlash may raise costs and delay permits for later-stage projects.
At SEMICON Taiwan 2026, TSMC announced an aggressive upgrade path for its COUPE (Compact Universal Photonic Engine). The bandwidth of a single COUPE is projected to double from 3.2Tbps to 6.4Tbps by 2027, with a future goal of exceeding 51.2Tbps. Additionally, equipment providers such as ficonTEC, Suruga Seiki, TOYO Automation, GMT Global, Allring Tech, and FitTech updated their respective progress on CPO (co-packaged optics) optical coupling and alignment equipment during SEMICON. This indicates that equipment vendors are accelerating their positioning to meet CPO mass production demands.
This report provides an in-depth analysis of (1) the FAU coupling in the CPO manufacturing process, (2) active versus passive alignment, (3) major suppliers of FAU coupling equipment, and (4) CPO market trends. The goal is to clarify the role of FAU coupling in CPO manufacturing, the competitive landscape among major equipment suppliers, and the latest CPO market trends.
NOR Flash memory is facing a structural imbalance as major international manufacturers reallocate capacity to high-end AI products, contracting supply amid surging demand from servers and automotive applications. Delayed capacity expansions and resource diversion among vendors mean high-density supply shortages remain difficult to bridge, driving prices continuously upward. This profitable cycle is expected to persist through next year and beyond.
AI servers power surging demand, yet U.S. grid capacity trails shipments as midterm politics raise deferral risk.
In the OCP APAC Summit held in August 2026, suppliers including TSMC, ASE, and AMAT all highlighted the future trends of 3D IC. In September, during SEMICON, the 3DIC Advanced Manufacturing Alliance (3DICAMA), initiated by TSMC and ASE, further organized the 3DIC Global Summit. The event shifted its focus toward mass production readiness, system integration, and ecosystem collaboration in the AI era, reflecting that 3D IC is entering a new phase of large-scale manufacturing and system-level optimization.
This report provides an in-depth analysis of: (1) the definition and technology types of 3D IC; (2) the Hybrid Bonding process; (3) a summary of Hybrid Bonding suppliers; (4) TSMC’s SoIC capacity expansion timeline and market dynamics; and (5) the systemic technical challenges and future trends of 3D IC. The goal is to explain the technical principles of Hybrid Bonding, review the major bonder suppliers and market landscape, and outline the future development direction of 3D IC.
New CEO Ternus leads Apple's first foldable launch; modest specs upgrade amid surging memory costs puts pricing and AI delivery at center stage.
On July 27, 2026, the Chinese AI startup Moonshot AI officially open-sourced the weights for its flagship AI model, Kimi K3. Boasting a total parameter count of 2.8 trillion (2.8T), Kimi K3 currently stands as the world’s largest open-source model, establishing a new benchmark for frontier AI models in China. Shortly after, DeepSeek released DeepSeek V4 Flash on July 31, and Alibaba Cloud announced Qwen 3.8 Max on August 3. These rapid developments have fueled market discussions regarding a potential “second DeepSeek moment” and the intensifying US-China AI rivalry.
This report provides an in-depth analysis of the following: (1) an overview of Kimi K3, Qwen 3.8 Max, and DeepSeek V4 Flash; (2) a dissection of Kimi K3’s core technologies; and (3) a comparative analysis of AI models developed in China and the United States. This report aims to evaluate mainstream Chinese AI models, elucidate the underlying technical principles and trends of Kimi K3, and assess the competitive landscape between Chinese and American AI capabilities.
The market once expected 5G to become the universal connectivity standard for the Internet of Things (IoT) and believed that AI would drive computing power toward the network edge. However, as of 2026, AI inference is instead shifting toward data centers and end devices, while connectivity technologies continue to diversify to meet the requirements of different applications.
Competition in the IoT chip industry is also evolving as technology roadmaps become increasingly fragmented, with new value gradually shifting toward platform solutions that integrate AI, multimode connectivity, and security. The report examines how the shift of AI inference toward endpoint devices, along with increasingly diversified connectivity, is reshaping the distribution of value and redefining competition in the IoT chip market.
This report focuses on the development trends of AI agent applications, revealing how they are changing the existing social landscape, and analyzes the possible shifts in AI core computing hardware configurations amid the rise of AI agent applications, as well as the supply chains expected to benefit directly.
