Recent #TSMC news in the semiconductor industry

4 months ago

➀ Ex-Intel CEO Pat Gelsinger advises Japan's Rapidus to develop unique differentiating technologies beyond production efficiency to compete with TSMC.

➁ Rapidus plans to integrate wafer fabrication and advanced packaging at the same facility for faster cycles, though full automation will not be available immediately from 2027.

➂ The company aims to begin 2nm test production with GAA transistors and establish a chiplets R&D center, utilizing ASML EUV lithography tools for future HBM and 3D packaging.

2nmAdvanced PackagingChipletEUVIntelRapidusTSMC
4 months ago
1. TSMC announces its A16 chip manufacturing technology, set to debut in 2026, using GAA transistors to enhance AI chip performance and power efficiency; 2. The technology will directly compete with Intel’s 14A process node, intensifying the semiconductor manufacturing rivalry; 3. GAA transistors enable higher transistor density and energy efficiency, critical for advancing AI and high-performance computing applications.
AI ChipsIntelTSMC
4 months ago
1. TSMC expands semiconductor production in the U.S. and Japan with government subsidies, facing challenges like construction delays and labor shortages in the U.S.; 2. Geopolitical tensions drive TSMC to diversify manufacturing bases to mitigate supply chain risks; 3. Japan's efficient subsidy distribution contrasts with U.S. bureaucratic hurdles, reflecting how state support shapes global chip industry dynamics.
GeopoliticsTSMC
4 months ago

➀ Counterpoint Research's analysis defines Foundry 2.0 as encompassing packaging and mask-making alongside wafer fabrication, with TSMC leading at 35% market share and Intel ranking second at 6.5%;

➁ The expanded definition may mitigate monopoly concerns for TSMC, whose traditional wafer fabbing dominance reaches nearly 70%;

➂ The Foundry 2.0 market reached $72 billion in Q1 2025, with Intel gaining 0.6% quarterly market share but declining 0.3% year-over-year.

IntelTSMCsemiconductor
5 months ago
1. TSMC delays production at its Arizona semiconductor plant to 2025 due to a shortage of skilled workers; 2. Disagreements with local unions over training and hiring practices further complicate the project; 3. The company plans to bring in temporary workers from Taiwan, sparking tensions over labor rights and expertise requirements.
TSMC
5 months ago
1. TSMC introduces its advanced A16 chip manufacturing process, targeting mass production by 2027; 2. The A16 technology integrates new materials and structural innovations to enhance transistor density and energy efficiency; 3. The process aims to address challenges in next-gen AI and high-performance computing chips.
TSMC
5 months ago
1. TSMC's SoIC advanced packaging technology faces explosive demand driven by AI chips and HPC applications; 2. Major OSAT companies like ASE and Amkor are accelerating capacity expansion, particularly for CoWoS and chip-on-wafer packaging; 3. The supply chain faces bottlenecks as equipment delivery times extend to 18-20 months, prompting strategic collaborations across the ecosystem.
Advanced PackagingTSMC
5 months ago
1. TSMC's 3nm semiconductor manufacturing process is reportedly facing yield challenges, particularly with the N3E node; 2. The issues may impact production timelines for Apple's upcoming M4 and A18 chips, slated for 2024 devices; 3. Despite improvements in N3P node yields, supply constraints could affect Apple's product roadmap and TSMC's competitiveness against Samsung Foundry.
3nm processTSMC
5 months ago
1. TSMC secured $6.6 billion in U.S. subsidies under the CHIPS Act to build advanced semiconductor plants in Arizona; 2. The funding supports three fabrication facilities, boosting U.S. chip production and creating thousands of jobs; 3. The move aligns with U.S. efforts to reduce reliance on Asian chip manufacturing and counter China's tech ambitions.
IntelTSMC
5 months ago

➀ Tesla developed a groundbreaking 'Stress' tool to detect defective cores in its wafer-scale Dojo processors, critical for preventing silent data corruption that could invalidate weeks of AI training;

➁ Each Dojo 'Training Tile' contains 8,850 RISC-V cores and leverages TSMC's InFO_SoW packaging, with Stress enabling real-time monitoring across millions of cores without downtime;

➂ The method not only identifies faulty cores but also exposed rare design flaws, positioning Tesla alongside Google and Meta in hardware reliability while shaping future wafer-scale chip adoption.

AI ChipTSMCTesla