ByteDance is in talks to buy AI chips from Shanghai-based Iluvatar CoreX and is also considering using Baidu's Kunlunxin chips, as the TikTok parent expands its domestic chip options amid rising inference demand from its AI chatbot Doubao.
Apple is stepping up the AI capabilities of its Siri voice assistant, and analysts say memory chip demand will rise along with it, potentially benefiting Apple's suppliers such as Samsung Electronics and SK Hynix. The shift could drive both shipment growth and higher prices for mobile DRAM.
China has recently eased controls on some indium phosphide (InP) substrates, relieving a bottleneck in optical communications capacity for the second half of the year. But supply chain players say the long-term priority is still to expand substrate capacity from non-China sources, with supply security for the AI industry outweighing price.
Nvidia's dominance in AI is extending beyond model training and deeper into inference—the fast-growing segment of the AI market responsible for running deployed models and generating revenue.
Physical AI is emerging as a new frontier of model development. Any model, however, is only as good as the data used to train it. Because of this, Japanese startup APTO is launching a physical AI infrastructure lab to help plug the data gap needed to create vision-language-action (VLA) models, with a focus on imitation learning.
Samsung Electronics' foundry division chief told employees on June 12 that a return to profitability in the contract chipmaking business looks difficult next year, with 2028 emerging as a more likely timeline, Yonhap, ZDNet Korea, and Chosun reported.
Amid widespread anxiety sparked by foreign media reports about an impending "Tokenpocalypse," Apple software engineering chief Craig Federighi signaled that Apple is not pursuing "AI for AI's sake" at WWDC 2026, echoing a growing trend among Silicon Valley tech giants cautioning employees against using AI for the sake of using it.
Taiwanese companies sharply increased enterprise AI investment and adoption in 2026, yet critical gaps in technology architecture and measurable return on investment risk blunting business impact, according to Dun & Bradstreet's latest Enterprise AI Maturity Index. The index surveyed more than 300 Taiwanese firms across 17 industries as part of a global study of over 10,000 C-level executives in 32 advanced countries, finding momentum rising in the second quarter of 2026 but persistent operational hurdles.
Electromagnetic Compatibility (EMC) testing has historically been the bane of the power electronics industry. Engineers have long approached the compliance phase through trial and error, treating the process as a troublesome obstacle rather than a milestone. This reactive methodology commonly stalls product launches and drags out development, as teams scramble for solutions late in the manufacturing cycle. Adding to the pressure is a demographic shift occurring across modern engineering departments. As senior magnetic components experts retire, an influx of entry-level engineers is being tasked with navigating complex high-frequency board layouts.
China has allowed the release of a fresh supply of indium phosphide (InP) substrates, which are under export controls. A first 2026 batch shipped at the end of May following an earlier release in 2025, easing a capacity bottleneck in the optical communications market. Taiwanese compound semiconductor suppliers including Visual Photonics Epitaxy (VPEC) and Global Communication Semiconductors (GCS) are expected to benefit in the second half of 2026.
