Led by advanced packaging chief Hsu Kuo-chin, the world’s top logic foundry evaluates building vertical 3D DRAM fabs as AI GPU memory demand outstrips SK Hynix, Samsung, and Micron.
Featuring a doubled 2048-bit interface and TSMC-fabricated 3nm base dies, HBM4 becomes the most sought-after component in the global AI infrastructure buildout.
With Chip-on-Wafer-on-Substrate demand tripling over a 24-month window and test-equipment lead times stretching, foundry pricing power hits an all-time high.
Capitalizing on Nvidia’s leadership transition and strict allocation freezes, AMD and Broadcom lock in multi-billion-dollar cluster expansions with top cloud providers.
With Anthropic committing $705 billion across AWS Trainium and Google TPU clusters, hyperscaler custom chips move from internal experiments to primary revenue engines.
Washington moves to shut down the offshore cloud GPU rental pipeline that allowed Chinese AI developers to train frontier models on Nvidia B200 clusters hosted in Southeast Asia and the Middle East.
Jensen Huang resigned as CEO of Nvidia after 33 years effective immediately following the board’s probe into his relationship with Alice Huang and 240,000 diverted GPUs. CFO Colette Kress takes over as interim CEO.
A forensic study of public model weight initializations traced distinct floating-point signatures in DeepSeek, Moonshot AI, and MiniMax models directly to restricted Nvidia Blackwell B200 and GB200 clusters.
As Jensen Huang admits to a multi-year affair with suspected Chinese intelligence asset Alice Huang, US lawmakers, the FBI and Nvidia’s board are investigating a burning question: did she use her personal access to the Nvidia chief to secure $7.5B in banned H100 and H200 GPUs?