Reports
AI-generated structured vendor updates
NVIDIA Employee Detained in Taiwan B300 Smuggling Probe, Export Control Reaches Chipmaker
Taiwan prosecutors detain a NVIDIA employee for allegedly helping smuggle Super Micro servers with **GB300** chips to China. Seven people are held, involving $21 million. This marks the first time export control enforcement directly hits a chipmaker, shifting compliance risk upstream.
NVIDIA Vera Rubin NVL72 Enters Production, 10x Energy Efficiency Reshapes AI Infrastructure
NVIDIA has announced full production of the Vera Rubin NVL72 platform, with first shipments to CoreWeave, Microsoft, Amazon, and Oracle. Each rack integrates 72 Rubin GPUs and 36 Vera CPUs, delivering 10x improvement in energy efficiency and inference cost over Blackwell for agentic AI workloads, marking a new era of rack-scale AI infrastructure.
NVIDIA Prepay $1.5B to Amkor for US 2nm/3nm/HBM4 Advanced Packaging
NVIDIA prepays $1.5B to Amkor to expand advanced packaging capacity in Arizona, covering 2nm/3nm/HBM4. This onshores AI chip packaging, complementing Wistron's system integration, to reduce reliance on Taiwan and secure supply for Vera Rubin and Blackwell Ultra.
JCET's $1.4B AI Packaging Capex Reshapes Advanced Packaging Supply Chain
JCET announces $1.4B capex for AI advanced packaging in 2026, targeting Chiplet, HBM, 2.5D/3D. This marks Chinese OSAT's major entry into high-end packaging, complementing TSMC's CoWoS expansion and shifting AI packaging from Taiwan-centric to cross-strait division, supporting local hyperscalers' computing needs.
NVIDIA Denies Kyber NVL144 Delay, But 78-Layer PCB Bottleneck Exposes AI Hardware Physics Limit
NVIDIA officially denies reports of Kyber NVL144 rack delay to 2028, but SemiAnalysis revelations about a 78-layer ultra-high-density PCB midplane bottleneck and Rubin Ultra cancellation expose hard physical limits in signal integrity and manufacturing, opening a strategic window for AMD and Google.
AWS boosts Trainium 3 shipments, accelerating ASIC substitution for NVIDIA GPUs
Supply chain sources indicate Amazon AWS has instructed vendors to increase Trainium 3 shipments for Q3 2026 by 20-30%. This signals strong confidence in its custom ASIC strategy to reduce dependence on NVIDIA GPUs, leveraging superior cost and power efficiency for cloud AI training.
Samsung Restarts 1.4nm Foundry Node, Pre-emptively Locks Equipment Supply Chain
Samsung Electronics restarts 1.4nm (SF1.4) process commercialization, ordering equipment vendors to develop tools early. The node will use High-NA EUV lithography and GAA transistors, fabbed at NRD-K campus. This move aims to catch up with TSMC and Intel, but mass production timeline remains undisclosed.
TSMC Adds Winbond to WoW 3D Stacking Memory Supply, Breaking DRAM Oligopoly
Winbond joins TSMC's Wafer-on-Wafer (WoW) 3D stacking advanced packaging supply chain, becoming a new DRAM wafer supplier alongside Samsung, SK Hynix, and Micron. This move reduces reliance on the three global DRAM giants and strengthens AI chip packaging supply resilience. Winbond provides DRAM wafers for vertical stacking with TSMC logic wafers, offering 8GB capacity and 256GB/s bandwidth via its CUBE solution.
TSMC Bets on CoPoS and Glass Substrates: Packaging Paradigm Shifts from Wafer-Level to Panel-Level, AI Chip TCO Inflection
TSMC is replacing CoWoS with CoPoS (panel-level packaging), using 750x620mm square panels and glass core substrates, achieving 20-30% unit area cost reduction. Volume production targets 2028, with AMD Zen 7 as first key customer. This fundamentally alters AI chip packaging economics and capacity scaling.
MediaTek Pivots to System-Level Integration: Targeting Google TPU and Musk AI Rack Deals
MediaTek elevates its AI strategy from chip design to system-level integration, targeting Google TPU PCBA L6 and Musk AI chip L10 rack assembly. Adopting a light-asset model via Taiwan's supply chain, targeting >40% gross margin, driven by rising complexity from CPO and 800V high-voltage DC power.
MediaTek Shifts from Chip Design to System Integration for Google TPU and Musk's AI Racks
MediaTek upgrades its AI strategy from chip/ASIC design to system-level integration, targeting Google TPU PCBA and Musk's AI chip rack-level (L10) work. Using an asset-light model, it leads design and validation while outsourcing manufacturing, aiming for 40-50% gross margin.
MediaTek Pivots from Chip Design to System-Level Integration, Targeting Google TPU and Musk AI Racks
MediaTek elevates its AI strategy from chip design to system-level integration, targeting Google TPU v10 PCBA and Musk-affiliated AI rack assembly. Using an asset-light model and Taiwan's supply chain, it aims for 40-50% gross margin in system integration.
NVIDIA Vera 88-Core Arm CPU: Control Plane Shifts from x86 to NVIDIA for AI Agent Workloads
NVIDIA unveils Vera, its first standalone datacenter CPU with 88 custom Arm Olympus cores, monolithic mesh, 1.2TB/s LPDDR5X bandwidth, achieving 1.8x x86 performance in agent workloads. Tightly coupled with GPUs via NVLink-C2C, Vera shifts the control plane from Intel/AMD to NVIDIA. First customers: OpenAI, Anthropic. Production Q3 2026.
TSMC 2026 Outlook: AI Demand Drives 30%+ Revenue Growth, Advanced Process and Packaging Dual Constraints
Behind TSMC's revenue growth forecast is dual logic of 'volume and price both rising': AI chip demand drives shipment growth, advanced process scarcity pushes wafer unit prices up. But A16 process delay is a signal worth watching—even TSMC faces increasing difficulty in advanced process mass production.
Intel Foundry Breakthrough: EMIB Packaging Gains Strategic Endorsement from Google, Amazon
The strategic significance of this deal far exceeds surface numbers. Google's and Amazon's simultaneous shift to Intel signals: US cloud giants' strategic consensus on 'de-TSMC-ization' in AI chips has formed. Not just chip manufacturing, but advanced packaging—high-value-added manufacturing—is also undergoing supply chain restructuring.
Google Upgrades Medical Workflows with AI Agent and Health Data Integration
Google invests $10M in AI training for clinicians and integrates Gemini AI with 20 years of health data for predictive diabetes care and automated patient experiences. This validates AI's transition from辅助工具 to core agent in medical workflows.
Ericsson Deepens 5G Advanced and AI-Native Network Collaboration with FarEasTone
Ericsson signs a three-year strategic agreement with FarEasTone to advance 5G Advanced deployment and AI-native network R&D. The collaboration includes key technologies like network slicing, URLLC, and RedCap, enabling Industrial 4.0 and smart city applications. Both parties will jointly research 6G and AI-driven network automation.
Google Builds Diabetes Prediction Blueprint with 20-Year Health Data and Gemini
Google partners with Taiwan to develop a diabetes prediction system using 20 years of health data and Gemini AI. The project demonstrates a blueprint for combining generative AI with large-scale real-world data in public health. The technical core involves improving chronic disease prediction accuracy through long-term data training.
Google Collaborates with Taiwan on Gemini-Powered Public Health AI Model
Google partnered with Taiwan's health authority to develop AI-on-DM diabetes risk assessment model, using Google Cloud's concurrent processing to reduce evaluation time from 20 minutes to 25 seconds. Gemini health assistant integrated into official health app provides personalized recommendations, with framework expandable to other chronic diseases.
Google Launches Festival AI Music Generation Feature Based on Lyria 3 Model
Google introduced a limited-time AI music creation feature in Gemini app, using Lyria 3 model to generate personalized music cards. The feature allows multi-region users to create 30-second hi-fi audio and custom covers through prompts, with social sharing integration.