Global semi news — Korea, China, Taiwan, the US, and Japan. Government policy, export controls, capex moves, supply-chain shifts, and macro events. AI-classified and tagged with affected tickers. All headlines link back to the originating publisher.
AMD CEO Lisa Su said TSMC's Arizona fab ramp is going well, with AMD already taping out and producing chips there. The endorsement reinforces TSMC's US capacity buildout as a viable second source for leading-edge logic amid ongoing tariff and geopolitical pressure on Taiwan-only production.
Why it matters: Positive directional update on TSMC's Arizona ramp from a key customer, but no new capex figures or specific guidance changes.
Original: Taiwan probes alleged smuggling of Nvidia AI chips to China in possible export control breach - Malay Mail
Taiwanese authorities have opened an investigation into the suspected smuggling of Nvidia AI accelerators into China, a potential violation of US export controls. The probe adds to mounting scrutiny of diversion channels through Taiwan and Southeast Asia, and could prompt tighter end-use checks on NVDA shipments and on Taiwanese assembly/logistics partners.
Why it matters: Direct export-control enforcement action targeting Nvidia AI chip diversion through Taiwan — material for NVDA and Taiwanese supply-chain partners (TSMC, Hon Hai, ASE) given the policy escalation risk.
Original: US pumps $2bn into quantum computing via CHIPS Act - Silicon Republic
The US government is allocating roughly $2bn from CHIPS Act funds to accelerate domestic quantum computing R&D and manufacturing. The move broadens CHIPS Act spending beyond classical logic/memory fabs, signaling sustained federal support for advanced compute infrastructure that ultimately relies on leading-edge foundry and packaging capacity.
Why it matters: Sector-wide US policy signal expanding CHIPS Act scope to quantum; no direct allocation to tracked KR/TW/US names, but advanced foundry and packaging players are indirect beneficiaries.
Open source articleOriginal: Why Nvidia stock is barely moving after earnings crushed expectations - MSN
Nvidia delivered another earnings beat that topped consensus, yet shares are showing a muted reaction as expectations were already elevated heading into the print. The piece probes why the upside surprise failed to drive a fresh leg higher, suggesting the AI-chip narrative is now largely priced in.
Why it matters: NVIDIA earnings and post-print reaction are a direct read-through for the entire AI semi supply chain including HBM, foundry and ASIC peers.
Original: Chip Industry Week In Review
Weekly digest covering Taiwan and Europe advanced packaging capacity additions, 2nm production ramps, major quantum funding, and an averted Samsung labor strike. Also notes a quantum-dot qubit fabricated with high-NA EUV and an EU flagship power electronics project — broad sector touchpoints but no single company-moving event.
Why it matters: Weekly roundup touches sector-wide themes (advanced packaging capacity, 2nm ramp, Samsung labor) relevant to peers but lacks a single near-term company-specific catalyst.
Original: 엔비디아 베라 CPU, 컴퓨텍스 2026서 인텔·AMD x86 대비 1.5배 성능 전망
Analysts project NVIDIA's upcoming Vera CPUs, expected to be showcased at Computex 2026, will outperform Intel and AMD x86 chips by roughly 1.5x. The Arm-based Vera is part of NVIDIA's Vera Rubin platform and signals deeper CPU competition for the data center, pressuring x86 incumbents while reinforcing Arm's AI infrastructure footprint.
Why it matters: Pre-event analyst projection on NVIDIA's next-gen Arm CPU vs x86 incumbents — a sector-wide AI infra/CPU competitive theme rather than a confirmed product event.
Open source articleOriginal: Arm, 엔비디아 AI CPU 확장에 힘입어 16% 급등하며 사상 최고가 경신
Arm Holdings (ARM) jumped 16% to an all-time high after Nvidia's expanded AI CPU push reinforced Arm's positioning as the core architecture for next-gen data center compute. The move strengthens Arm's royalty growth narrative as Nvidia scales Grace-based and custom Arm-based CPUs across AI infrastructure deployments.
Why it matters: Direct stock-moving event on a tracked name (ARM) with read-through to Nvidia's AI CPU roadmap.
Open source articleOriginal: AMD, AI CPU 수요가 전망치 초과하자 생산 확대
AMD is accelerating production to meet AI CPU demand that is exceeding internal forecasts, signaling sustained tightness in AI server silicon. The ramp implies incremental orders for foundry partner TSMC and advanced-packaging/HBM suppliers feeding AMD's MI-class and EPYC roadmaps.
Why it matters: Direct production ramp by AMD driven by AI CPU demand is a hard datapoint for AMD, TSMC, and HBM/packaging suppliers.
Original: 엔비디아 vs 인텔, COMPUTEX서 AI CPU 정면 승부
Nvidia and Intel are set to face off at COMPUTEX with competing AI-oriented CPU offerings, signaling an intensifying battle for the next-generation AI client and edge compute market. The showdown highlights Intel's push to defend x86 against Nvidia's expanding silicon roadmap, with implications for the broader AI infrastructure stack.
Why it matters: Preview of a major CPU competition reveal at COMPUTEX directly involves two tracked names (NVDA, INTC) but lacks concrete product specs or earnings impact.
Open source articleOriginal: 모건스탠리, 엔비디아 베라루빈 랙 단가 780만달러 추정
Morgan Stanley estimates a single Nvidia Vera Rubin NVL144 rack costs north of $7.8M, up sharply from Blackwell-generation racks. The higher BOM reflects richer HBM content, CoWoS packaging, and networking silicon per rack, signaling continued AI capex intensity for hyperscalers and ASP tailwinds for the Nvidia supply chain.
Why it matters: Sell-side BOM estimate for Nvidia's next-gen rack reinforces ASP and HBM/packaging content uplift across the Nvidia supply chain.
Kioxia
285A
¥67,100
-12.86%