NVDA•NVDA•SK Hynix plans a $28B U.S. IPO to tap AI-driven memory demand, potentially improving Nvidia’s DRAM supply outlook, while Foxconn’s Q2 revenue jumped 39.8% to NT$2.513T despite geopolitical caution. Meanwhile, Nvidia moves to monetize AI compute by launching a startup-focused GPU credit program as Chinese IC design revenues near CNY1T.
SK Hynix filed for a $28 billion initial public offering in the U.S. to capitalize on surging AI-driven demand for high-bandwidth memory, a move that could bolster memory availability and pricing dynamics critical to Nvidia’s GPU production.
Foxconn reported April–June revenue of NT$2.513 trillion, up 39.8% year-over-year, driven by strong device assembly orders, but warned that escalating U.S.-China tensions could disrupt its customer supply chains, including for Nvidia components.
China’s domestic IC design sector is on track to exceed CNY1 trillion (≈US$150 billion) in annual revenues ahead of schedule, yet local chipmakers still trail Nvidia’s CUDA ecosystem, underscoring a persistent architecture and talent gap.
Self-driving startup Turing secured backing from AMD and announced plans to integrate AMD’s MI300 GPUs into its autonomous vehicle compute stacks, signaling intensifying competition in the AI accelerator market that Nvidia dominates.
Nvidia is rolling out a new AI compute credit program aimed at early-stage companies, offering on-demand access to H100 and A100 GPUs through prepaid credits at premium rates to capture a larger share of emerging AI workloads.
