As demand strengthens in markets such as GPUs and high-bandwidth memory (HBM), Japan-based Advantest said orders for semiconductor automated test equipment (ATE) remain stable, with average product lead times exceeding six months. The company expects that once the newly expanded assembly capacity is fully in place in 2026, order backlogs will be further alleviated.
OpenAI has reached a multi-year compute procurement agreement worth more than US$10 billion with AI chip startup Cerebras Systems, with the intention to enhance ChatGPT performance and reduce its reliance on Nvidia. According to The Wall Street Journal and Bloomberg, OpenAI said in a statement that it has committed to purchasing up to 750 MW of computing capacity from Cerebras by 2028. Contract terms were not disclosed, but people familiar with the matter said the deal value exceeds US$10 billion.
In August 2024, the US National Institute of Standards and Technology (NIST) finalized three Post-Quantum Cryptography (PQC) standards, known as FIPS 203, 204, and 205. These standards incorporate ML-KEM algorithms for key exchange and encryption, along with ML-DSA and SLH-DSA algorithms for digital signatures. Industry leaders are actively integrating these algorithms into web browsers, operating systems, and hardware products to prepare defenses against anticipated quantum-computer-enabled cyberattacks.
As artificial intelligence integration accelerates within enterprises, the cybersecurity landscape is undergoing significant change, with machine identities now surpassing human users in many organizations. This shift, highlighted by Simon Tai, managing director of SailPoint Taiwan, Hong Kong, and Macau, signals the growing inadequacy of traditional governance frameworks for effectively securing AI-driven operations.
OpenAI has underscored the increasing complexity and unpredictability of the AI rivalry between the US and China, highlighting China's rapid progress through open-source, low-cost AI models and its export of "sovereign AI" systems to Southeast Asia and the Middle East. These shifts indicate new geopolitical dimensions in the global AI landscape.
Touch panel manufacturer Young Fast Optoelectronics, which has shifted focus toward electromechanical products in recent years, plans to vertically integrate its optoelectronics division and introduce two new business segments by 2026. The company expects to boost growth by adding full assembly for gaming machines and general displays, as well as medium-to-large electronic paper (EPD) module lamination, with visible results anticipated by late first quarter 2026.
Apple has officially confirmed a strategic collaboration with Google to use the Gemini AI model to enhance Siri and other Apple Intelligence features. The two tech giants have revealed limited details about the arrangement, with internal sources shedding light on key aspects of the cooperation that carry broad implications for Apple's AI development and competitive positioning.
Humanoid robots and embodied intelligence were huge at CES 2026. Dexterous hands are a key component that determines the upper limits of a robot's manipulation capabilities. Despite significant progress in recent years in terms of degrees of freedom and control performance, several challenges remain to be addressed.
Meta Green Cooling (MGC) is making a comprehensive push into the AI liquid cooling sector, leveraging its unique full-spectrum capabilities. MGC chairman James Chen noted that, unlike competitors, MGC integrates mechanical design, electromechanical systems, thermal management, monitoring, data center design, and maintenance services under one roof. The company posted a remarkable 215% revenue growth in 2025 and expects similar expansion in 2026 as it capitalizes on rising liquid cooling demand.
AMD introduced its Helios system platform at CES 2026, marking a significant move in cloud AI computing competition from traditional single-chip performance to full rack-scale solutions. This development follows Nvidia's GB200 NVL series release and Google's TPU-based rack product plans disclosed by Broadcom, underscoring an industry trend toward delivering computing power through integrated system platforms rather than isolated chips.
Render Network forecasts that 2026 will see brands adopting fully autonomous AI systems that reflect their identity across customer interactions, while consumer GPUs will enable creators and studios to run advanced AI models locally, expanding access to powerful AI workflows beyond large enterprises.
More coverage
