IT之家

Headlines · 5

1. Oct 20
   TrendForce sees Q4 2026 DRAM contract prices up about 20% on AI inference demand

   TrendForce estimates that DRAM contract prices will rise about 20% quarter-over-quarter in the fourth quarter of 2026, according to IT之家 and a 36氪 brief citing 财联社. That pace is slower than the roughly 80% increase the research firm cites for the second quarter and about 40% for the third quarter, though DRAM prices remain high in a cycle of tight supply and strong demand. IT之家 attributes the increase mainly to expanding AI inference demand, with larger models, longer context windows and growing KV cache needs pulling up server DRAM and enterprise SSD capacity and bandwidth requirements while memory makers prioritize high-end products such as HBM. The slower pace is put down to an already high price base, weaker end-market purchasing power and price ceilings in some long-term procurement agreements; IT之家 also reports that Nvidia is considering cutting Rubin Ultra's HBM stack count from 12 to 8 layers due to tight HBM supply.

   IT之家 AI · 🔥 0
2. Oct 19
   Dongfeng Mengshi X700 set for Oct 25 launch with Huawei Qiankun ADS 5

   IT之家 reported on Oct 19, citing Yiche, that the Mengshi X700 will go on sale on Oct 25 in three variants with presale prices starting at 249,800 yuan. It is the first product in Mengshi's X series, opened for presale on Sept 28, and carries Huawei Qiankun ADS 5 driver assistance plus the HarmonySpace 6 cockpit. The report lists an 896-line roof lidar, highway and city NOA, memory parking, AEB 3.0 and point-to-point parking, along with range-extender and plug-in hybrid powertrains, an 800V platform and a 68.1kWh CATL off-road battery with 360km CLTC electric range.

   IT之家 AI · 🔥 0
3. Oct 19
   MetaX forecasts RMB 587M–717M profit for first nine months of 2026

   Chinese GPU maker MetaX issued a preannouncement for the first three quarters of 2026 (January–September), projecting revenue of RMB 2.541–3.106 billion, up 105.6%–151.29% year on year, and net profit attributable to shareholders of RMB 587–717 million versus a net loss of RMB 346 million a year earlier, according to IT之家. Non-recurring-adjusted net profit is expected at RMB 223–302 million, compared with a RMB 393 million loss last year. The company attributed the growth to its "1+6+X" strategy and market expansion, saying customers continued to buy its high-performance GPUs and services, with higher revenue plus increased investment in industry-chain companies driving the turnaround. MetaX cautioned that the figures are preliminary, unaudited, and that final numbers will come in its formal quarterly report.

   IT之家 AI · 🔥 0
4. Oct 18
   Intel calls silicon photonics next-gen 'Golden Hammer', claims sole 300mm InP volume fab

   At the 2026 OCP Global Summit, Intel data center systems vice president Nic Dubé described silicon photonics as the "Golden Hammer" of next-generation system architecture, saying it will accelerate high-performance segments including HPC and AI. Intel claims it is the only high-volume fab able to supply 300mm (12-inch) indium phosphide silicon photonics at scale, citing a unique track record in integrated laser reliability, and says it is pushing community standards for external light sources. Intel estimates photonics systems can reach transmission energy below roughly 5 picojoules per bit (IT之家 reported under 5.5 pJ/bit), and says it will offer advanced interconnect topologies with runtime orchestration of CPUs, GPUs and memory pools as AI models grow. The exclusivity and efficiency figures are Intel's own claims and were not independently verified in the reports.

   Wccftech · 🔥 0
5. Oct 16
   TSMC starts mass production of 5x-reticle advanced packaging

   TSMC Chairman and CEO C.C. Wei said on the company's fiscal third-quarter earnings call that TSMC has begun mass production of 5x-reticle advanced packaging with “very excellent” yields, claiming no other company can currently match that scale and that customers are satisfied with the results. He said TSMC is developing 9x-reticle packaging and pushing toward 14x; without naming Intel's EMIB or other competing approaches, he said TSMC would be ready if customers cannot secure enough advanced packaging capacity from competitors. IT之家's explainer notes that AI and high-performance computing chips increasingly rely on packaging multiple compute dies and high-bandwidth memory on one large substrate, so larger package areas usually allow more compute, memory and interconnect density but make yields harder to guarantee; the report separately cites Anue's August report that 5.5x-reticle CoWoS is in high-volume production with some customer yields above 98%, and that TSMC plans to extend CoWoS beyond 14x by 2029.

   IT之家 AI · 🔥 0

Experience and discussion from the community

Share my IT之家 experienceAsk about IT之家