Apple and Google have entered an AI technology partnership reportedly worth as much as US$5 billion, according to industry sources cited by the Financial Times. Meanwhile, Bloomberg reported that Apple may pay roughly US$1 billion annually to use a customized Gemini model developed by Google, which runs on Apple's own servers to maintain user privacy.
The escalating global energy transition and intensifying industrial rivalry have elevated power infrastructure security to a national defense-level strategic concern, with a focus on energy transmission efficiency rather than mere generation capacity. Experts warn that bottlenecks in ultra-high-voltage (UHV) transmission, particularly in 345 kV cables, are now decisive factors shaping industrial competitiveness and the global digital economy landscape.
Industry speculation about collaboration between optics manufacturers and semiconductor companies to bolster their silicon photonics (SiPh) capabilities is gaining ground. Following Largan Technology chairman Enping Lin's comments during a recent earnings call, Asia Optical chairman I-Jen Lai confirmed that Asia Optical is also a potential participant in this trend.
OpenAI said it would partner with the artificial intelligence chip maker Cerebras to build 750 megawatts of ultra-low-latency AI computing capacity, a project that will come online in phases starting in 2026. Construction is expected to begin this year and continue through 2028.
Wistron posted a record-high consolidated revenue of NT$2.19 trillion (approx. US$67.63 billion) for full-year 2025, according to its unaudited financial results announced on January 16. The company reported gross profit of NT$134.02 billion for 2025, while operating profit came in at NT$78.56 billion, income before tax totaled NT$76.26 billion, and net income after tax reached NT$27.401 billion.
Acer and Lenovo are optimistic about India's consumer electronics sector as the global market slows due to rising memory prices.
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.
More coverage
