The five largest cloud operators spent US$181.5 billion on property and equipment in the June quarter, 87% more than a year earlier. Nvidia's hyperscale revenue was US$48.7 billion — 26.8 cents of every dollar they spent, down from a peak of 33.2 cents three quarters ago.

Doosan Corporation's Electro-Materials BG has developed copper-clad laminate (CCL) technology that reduces the prepreg (PPG) insulating layer used in memory semiconductor package substrates from about 18µm to 13µm, and is preparing the material for mass production as thinner package designs raise the technical demands on substrate materials.

Kioxia Holdings is weighing construction of a new manufacturing building at its Kitakami plant in northern Japan, as rising NAND demand pushes the memory maker toward its next phase of capacity expansion beyond 2029.

South Korean memory rivals Samsung Electronics and SK Hynix are taking fundamentally different routes to solve HBM performance bottlenecks. While Samsung frames processing-in-memory (PIM) as its core solution, SK Hynix is prioritizing packaging interconnect innovation to streamline data movement.

Nvidia said in its fiscal 2027 second-quarter results on the 27th that it sold a small number of H200 artificial intelligence (AI) processors to Chinese customers last quarter, marking its first renewed AI chip sales to China since shipping about US$60 million of H20 chips in early 2025. But shipments fell short of the total approved by US President Donald Trump and licenses cleared by the US government in January, amid opposition from Beijing.

Nvidia closed its second fiscal quarter with US$31.6 billion of inventory, but the more consequential change is what that inventory now consists of. Work in process and raw materials together account for 78% of the balance, against 42% a year earlier. The mix shows a company committing to scarce memory well ahead of the Vera Rubin ramp, and absorbing the cost several quarters before customer price increases land.

Nvidia's data center business now splits almost evenly between hyperscalers and everyone else — US$48.7 billion against US$40.3 billion in the July quarter. The obvious reading is that the customer base is broadening. The sequence says something more specific.

CXMT is narrowing the performance gap with SK Hynix in high-speed DDR5 memory, with new benchmark tests showing CXMT chips running stably at 8,000 MT/s and delivering productivity and gaming performance within about 1% of SK Hynix's latest M-die.

China's smartphone and memory supply chains are moving closer together, with Xiaomi's new Xring O3 processor potentially providing CXMT with an early commercial entry into the LPDDR6 era.

The recent disclosure of SpaceX's partnership with Nvidia to deploy Vera CPUs as part of the architecture behind SpaceXAI's Starmind AI satellite was an outright showcase of where the two tech giants are decisively headed in the agentic AI era, as their collaboration deepens into one huge, interconnected ecosystem of orbital compute.

Powertech Technology (PTI) is preparing to move its fan-out panel-level packaging (FOPLP) technology into mass production by mid-2027. AMD and Broadcom are backing demand, and most of the planned capacity has already been booked, as the Taiwan-based OSAT supplier builds a second growth engine around advanced packaging and optical integration.