TSMC

Twitter

Facebook

Copy Link

Key Facts

Founded1987

Key PeopleC.C. Wei

HeadquartersHsinchu, Taiwan

Top News

US export rules greenlight H200 sales to China, but strict limits put American buyers first

The US Department of Commerce approved limited exports of Nvidia's H200 and AMD's MI325X AI chips to China under strict conditions. Chinese buyers can only receive up to 50% of units sold to US customers, and domestic demand must be fully met first. However, China's customs authorities have reportedly told agents the H200 is not permitted entry, creating uncertainty despite the formal approval.

Policy35 sources

US export rules greenlight H200 sales to China, but strict limits put American buyers first

Policy35 sources

AMD unveils Helios rack-scale AI system with 72 MI455X accelerators and 256-core EPYC Venice

AMD revealed its Helios rack-scale AI system at CES, featuring 72 Instinct MI455X accelerators with 31 TB of HBM4 memory and delivering 2.9 FP4 exaflops for AI inference. The system pairs with AMD's 256-core EPYC Venice CPUs and targets hyperscalers like OpenAI, xAI, and Meta. AMD also introduced the MI400-series family, including MI430X for HPC and MI440X for enterprise deployments.

Technology7 sources

AMD unveils Helios rack-scale AI system with 72 MI455X accelerators and 256-core EPYC Venice

Technology7 sources

Jensen Huang heads to Taiwan to launch Nvidia headquarters and host trillion-dollar dinner

Nvidia CEO Jensen Huang is visiting Taiwan this month to formally sign construction plans for the company's first global headquarters outside the US. The visit includes hosting the iconic 'trillion-dollar dinner' with key supply chain partners including TSMC, Foxconn, and Quanta as surging demand for AI chips intensifies pressure on the semiconductor ecosystem.

Business2 sources

Jensen Huang heads to Taiwan to launch Nvidia headquarters and host trillion-dollar dinner

Business2 sources

Nvidia's $20B Groq bet and Vera Rubin platform reveal how AI inference is splitting the GPU era

Nvidia CEO Jensen Huang outlined a major strategic shift at CES 2026, emphasizing serviceability and inference economics while announcing a $20 billion Groq licensing deal. The move signals the end of one-size-fits-all GPUs as AI inference workloads split into prefill and decode phases, with the Vera Rubin platform designed for modular maintenance and continuous operation in constrained power environments.

Technology3 sources

Nvidia's $20B Groq bet and Vera Rubin platform reveal how AI inference is splitting the GPU era

Technology3 sources

Nvidia scrambles to meet surging China demand for H200 chips as orders hit 2 million units

Chinese technology companies have ordered over 2 million H200 chips for 2026, but Nvidia holds only 700,000 units in stock. CEO Jensen Huang confirms demand is "very high" as the company works with TSMC to ramp up production while navigating U.S. export licensing and awaiting Beijing's approval for shipments.

Technology12 sources

Nvidia scrambles to meet surging China demand for H200 chips as orders hit 2 million units

Technology12 sources

Chinese AI chipmakers Biren and Iluvatar race to Hong Kong IPO amid chip funding frenzy

Shanghai-based Biren has kicked off bookbuilding for a Hong Kong IPO targeting up to $624 million, becoming the first mainland GPU developer to list in the city. Rival Iluvatar CoreX followed days later with its own filing, riding a wave of investor enthusiasm after Moore Threads and MetaX saw their shares surge over 400% and 693% respectively on debut. But widening losses and U.S. trade restrictions cast shadows over the sector's path to profitability.

Business2 sources

Chinese AI chipmakers Biren and Iluvatar race to Hong Kong IPO amid chip funding frenzy

Business2 sources

Rapidus unveils glass interposer tech to challenge TSMC's dominance in AI chips production

Japan's state-backed chipmaker Rapidus has developed a prototype glass interposer that could cut AI chip production costs significantly. The company plans mass production by 2028, using 600mm glass substrates that yield 10 times more interposers than traditional silicon wafers. This breakthrough positions Rapidus to compete directly with TSMC as Japan invests heavily to rebuild its semiconductor industry.

Technology3 sources

Rapidus unveils glass interposer tech to challenge TSMC's dominance in AI chips production

Technology3 sources

Taiwan opens Nvidia-powered cloud center to advance sovereign AI and cement role in global tech race

Taiwan President Lai Ching-te inaugurated a new cloud computing center in Tainan equipped with the Nano 4 supercomputer, powered by 1,760 Nvidia H200 chips and 144 Blackwell chips. The 15-megawatt facility marks Taiwan's shift from hardware manufacturing to becoming an AI island, advancing its sovereign AI ambitions. Analysts forecast rising investor confidence in Taiwan's entrenched position within AI supply chains.

Technology2 sources

Taiwan opens Nvidia-powered cloud center to advance sovereign AI and cement role in global tech race

Technology2 sources

US Launches Review of Nvidia H200 Chip Sales to China as Demand Surges and Production Ramps Up

The Trump administration has initiated an inter-agency review that could authorize the first shipments of Nvidia's H200 AI chips to China, with a 25% fee attached. The move comes as Chinese companies rush to place orders, prompting Nvidia to consider ramping up production capacity for the powerful processors.

Policy17 sources

US Launches Review of Nvidia H200 Chip Sales to China as Demand Surges and Production Ramps Up

Policy17 sources

Intel Emerges as Alternative to TSMC's Strained CoWoS Packaging as AI Demand Surges

Intel's EMIB and Foveros packaging technologies are gaining traction as alternatives to TSMC's capacity-constrained CoWoS packaging, with major tech companies exploring Intel's solutions amid surging AI chip demand.

Business3 sources

Intel Emerges as Alternative to TSMC's Strained CoWoS Packaging as AI Demand Surges

Business3 sources

Elon Musk Announces Ambitious Plan to Outproduce All AI Chip Makers Combined with Tesla's Custom Silicon

Tesla CEO Elon Musk claims the company will manufacture AI chips at higher volumes than all other chipmakers combined, with AI5 chips ready for production and AI6 in development. The ambitious plan includes building a 'TeraFab' facility and annual chip iterations to support Tesla's autonomous driving and robotics initiatives.

Technology2 sources

Elon Musk Announces Ambitious Plan to Outproduce All AI Chip Makers Combined with Tesla's Custom Silicon

Technology2 sources

AI Bubble Fears Trigger Global Semiconductor Selloff as SoftBank Plunges 40%

Growing investor concerns about an AI investment bubble have sparked a major selloff in global semiconductor stocks, with SoftBank losing $50 billion in market cap amid questions about AI valuations and spending sustainability.

Business3 sources

AI Bubble Fears Trigger Global Semiconductor Selloff as SoftBank Plunges 40%

Business3 sources

Google Opens Largest AI Hardware Engineering Hub Outside US in Taiwan

Google has established its biggest AI infrastructure hardware engineering center outside the United States in Taipei, Taiwan. The multidisciplinary hub will house hundreds of employees and aims to accelerate AI innovation while strengthening Taiwan's position as a global technology partner.

Technology3 sources

Google Opens Largest AI Hardware Engineering Hub Outside US in Taiwan

Technology3 sources

Tesla's Ambitious AI Chip Vision: Musk Targets 200 Billion Units Annually as Company Advances In-House Silicon Development

Elon Musk reveals Tesla's massive AI chip ambitions, seeking 100-200 billion chips annually while advancing its in-house AI5 and AI6 processors. The company may build its own fabs due to foundry timeline constraints.

Technology9 sources

Tesla's Ambitious AI Chip Vision: Musk Targets 200 Billion Units Annually as Company Advances In-House Silicon Development

Technology9 sources

AMD Unveils Zen 6 and Zen 7 CPU Roadmap: 2nm Process Technology and Enhanced AI Capabilities Coming in 2026-2027

AMD confirms its next-generation CPU architectures with Zen 6 launching in 2026 on TSMC's 2nm process and Zen 7 featuring advanced AI engines. The roadmap includes significant performance improvements and new AI capabilities for both desktop and server processors.

Technology5 sources

AMD Unveils Zen 6 and Zen 7 CPU Roadmap: 2nm Process Technology and Enhanced AI Capabilities Coming in 2026-2027

Technology5 sources

Prev 1…6 789 Next