SK Hynix plans a $28 billion US listing to expand AI-focused memory capacity, potentially easing Nvidia’s DRAM supply constraints. Meta is piloting an AI cloud service to sell idle GPU cycles while AMD’s backing of self-driving startup Turing spotlights growing competition to Nvidia’s compute leadership.
SK Hynix has filed for a $28 billion US share sale to accelerate production of high-bandwidth memory modules for AI servers, a move that could relieve DRAM shortages and improve supply availability for Nvidia’s next-generation GPUs.
Meta is building an AI cloud infrastructure offering enterprises access to unused GPU cycles in its data centers, a strategy that could drive incremental sales of Nvidia accelerators if the pilot scales to commercial production.
AMD has invested in self-driving startup Turing, which will deploy AMD GPUs in its autonomous vehicle platforms, underscoring intensifying competition to Nvidia’s entrenched position in AI inference and embedded compute.
China’s IC design industry is on track to hit CNY 1 trillion (approx. $150 billion) this year, yet persistent gaps in computing architecture and high-end talent reinforce Nvidia’s lead through its mature CUDA ecosystem.
Digitimes
