Lite-On Technology delivered record profitability in the second quarter of 2026, with gross margin reaching 27.2%, up 5.1% from a year earlier. President Anson Chiu said the company has secured two additional cloud service provider (CSP) customers in 2026 and remains highly confident that second-half performance will exceed the first half.
OpenAI has uncovered additional instances of autonomous AI agents escaping their intended testing environments as it broadens an investigation sparked by the high-profile Hugging Face security incident, according to Reuters. The newly discovered cases were reportedly limited in scope and, based on current findings, none of the agents are believed to have left OpenAI's internal network.
MediaTek said at its latest earnings call that its second-generation data-center ASIC product remains on schedule for mass production in 2028, and that the Intel embedded multi-die interconnect bridge (EMIB) advanced packaging process it uses has already reached a very good yield level. MediaTek also said its first-generation product remains on track for volume ramp-up in the fourth quarter of 2026 and is expected to generate US$2 billion in revenue.
IBM and Sarvam have announced a collaboration to advance sovereign AI technologies in India, a move that could influence how governments and regulated industries worldwide deploy trusted artificial intelligence. The effort focuses on secure, governed systems designed to support local control over data, operations, and compliance.
Token prices are falling while overall AI usage continues to surge — a trend this column examined in March 2026 and one that has only accelerated since. The next generation of frontier models is increasingly being built around the AI agent paradigm. In recent months, competition among leading AI labs has intensified, with successive releases demonstrating the ability to plan tasks, delegate work, use multiple tools and skills, and directly interact with computer interfaces — all at declining costs.
India is expanding its semiconductor ambitions through new R&D centers, indigenous design tools, packaging facilities and AI infrastructure investments, aiming to strengthen domestic capabilities and reshape global technology supply chains.
Agentic AI is reshaping how work is delegated among intelligent systems, yet legal frameworks have failed to keep pace with the growing autonomy of AI agents. Legal loopholes involving responsibility and authority must be addressed urgently, South Korean legal scholar Hyesun Yoon warned at an academic forum hosted by Taiwan's National Science and Technology Council (NSTC) on July 30.
The AI boom has been lifting Taiwan's upstream, midstream and downstream supply chains, but the country's application development remains a key area that still needs work. The Ministry of Digital Affairs (MODA) said Taiwan must accelerate the development of AI applications, services and software ecosystems, in addition to its existing strengths in semiconductors and hardware supply chains, if it wants to build a more favorable environment for AI industry growth.
For three years, the constraint on artificial intelligence was how much compute the hyperscalers wanted to buy. The June-quarter filings of seven companies say the constraint has moved. It is now about who sets the price — and on that question the past five quarters have already taught a lesson from which most readers drew the wrong conclusion.
The Monte Jade Science and Technology Association of Taiwan (MJ Taiwan) elected Maverick Shih, chairman of Acer Cyber Security (ACSI), as its 13th chairman during the association's first board meeting of its new term on July 30.
US transportation authorities have allowed Amazon's Zoox to operate commercially on a limited basis, making it the first paid robotaxi service approved in the country. This follows efforts by Trump administration officials to relax certain rules in an effort to encourage development in autonomous driving technology.
