Topic
Explore English articles about AI Semiconductor.
东方算芯's DF1000 stacks DRAM directly onto 14nm logic, claiming 6.4TB/s of bandwidth. We examine what this design means for moving away from HBM dependence, and the conditions that remain unverified for mass production and real-world measurement.
In July 2026, Moonshot AI's Kimi K3 took the top spot in the Arena coding rankings—but behind the scenes, it emerged that US-based Cursor had not initially disclosed its reliance on an earlier version of Kimi. This piece traces how suspicion of and dependence on Chinese-made AI became entangled over the course of just a few months.
Meta and Anthropic are in early talks over lending compute resources worth up to $10 billion over two years. If the deal materializes, it would add external sales as a new avenue for recouping Meta's massive AI investments.
Reports indicate that CXMT's DRAM orders are booked through the end of 2027. Reading the background of major PC manufacturer prioritization and the timeline for supply improvement, based on a 95.73% utilization rate and equipment upgrade plans through 2028.
In disputes over US memory patents, Kioxia was ordered to pay massive damages for infringing a Viasat patent, while Samsung and others face an import ban petition from Netlist. Not the memory cells themselves, but peripheral technologies like error correction and stacking control are having a major impact on product distribution.
Administrative records confirm that a memory factory operated by China's SwaySure is running, but the monthly capacity figures that have circulated for some time are not new information. While the US has added the company to its export control list—suggesting a close relationship with Huawei—significant technical barriers to HBM mass production remain.
The EU has required Google to share search data and open up AI features in Android 18. Combining anonymization, eligibility requirements, independent audits, and fee calculations, the move will reshape competitive conditions in search and mobile AI in phases starting in 2027.
Intel has planned a company-wide rollout of Gemini Enterprise and the parallelization of undisclosed types of HPC simulations using C4/N4. This is a collaboration that separates the roles of AI agents and semiconductor development.
Taiwan's Formosa Plastics will raise prices for electronic-grade hydrofluoric acid and other chemicals in Q3 2026, driven by growing demand from advanced semiconductor production increases and soaring raw material costs. While the company is also expanding supply capacity, the new capacity won't come online until 2027, raising concerns about continued supply-demand tightness in the near term.
OpenAI and Work Louder have opened preorders for a $230 desktop controller that moves status displays and approval actions for multiple Codex agents onto physical keys. We examine how it differs from existing products and the practical conditions involved.
