Topic
Explore English articles about AI Semiconductor.
The MCP 2026-07-28 specification eliminates connection-time initialization and session IDs, carrying version and capability information per request instead. While this eases horizontal scaling on HTTP infrastructure, it requires authorizing state handles and testing migration between old and new versions.
Gartner forecasts global AI spending will reach $2.59 trillion in 2026. Drawing on primary sources, this article examines how those costs are flowing through to Microsoft 365 price increases, cloud usage-based billing, and device prices.
Qualcomm has already publicly disclosed AI Frame Fusion, but internal documents reportedly claim it will be exclusive to the next Snapdragon Pro. This piece separates confirmed official technology from unverified product differentiation, examining implementation requirements and undisclosed benchmark figures.
Micron and Meta used DCPerf to examine how LPDDR5X speed and capacity affect AI preprocessing, web response, and Spark SQL. Running out of capacity and spilling to SSD produced the largest performance gap.
Samsung and Broadcom have signed an MOU covering HBM, sub-2nm manufacturing, and advanced packaging. The figure of over $200 billion is not a confirmed order, and disclosure of breakdowns and mass production conditions is needed for it to become real demand.
SK and NVIDIA have combined AI factory construction with long-term HBM4 supply into an LOI exceeding $500 billion. Initial operation of up to 2GW is planned for 2027, but the cost breakdown and construction site remain undisclosed.
Micron has opposed Apple's proposal to source memory chips made in China. A gap exists between near-term price relief and supply improvements expected from 2028 onward, with the Trump administration's additional regulations set to determine adoption conditions.
China's DRAM giant CXMT has set prices above Samsung's for 64GB DDR5 server memory and rejected even Huawei's demands for a discount. As AI-driven HBM production shifts squeeze general-purpose DRAM supply, this move signals a shift toward whoever holds pricing power.
Etched announced a $300M raise that included participation from SK hynix, and has begun initial production of inference clusters combining low-voltage compute with shared memory. Supply agreements and comparable performance figures have not yet been disclosed.
Reports suggest Intel could handle wafer manufacturing and advanced packaging for NVIDIA's next-generation AI platform Feynman. We examine what's confirmed in official materials versus what remains an unsettled contract scope.
