➀ Traditional RF board design strategies are insufficient for high frequencies and dense layouts;
➁ Keysight and Modelithics collaborate on an advanced RF board simulation workflow;
➂ The use of 3D passive component models enhances simulation accuracy and reduces the risk of board re-spins.
Genesem, a fab equipment manufacturer, has supplied a package sorter for the final testing of high-bandwidth memory (HBM) to a South Korean chip-maker. This is Genesem's first deal of this kind with the chip-maker. The sorter is also compatible with HBM inspection equipment that Nvidia requested.
➀ Samsung's top chip executive visited Nvidia's headquarters in the US;
➁ He brought a sample of the company's latest 1b DRAM;
➂ The purpose of the visit was to discuss the potential use of the improved DRAM for HBM.
➀ The GeForce RTX 5060's launch has been pushed back to April 2025 due to 'chip supply constraints';
➁ The delay affects the RTX 5070, RTX 5080, and RTX 5090, with limited supply expected leading to instant sellouts;
➂ The AI boom is increasing semiconductor demand, impacting the availability of the new GeForce RTX 50 Series.
➀ Anduril Industries, led by Palmer Luckey, is set to take over the US Army's Integrated Visual Augmentation System (IVAS) program, pending Department of Defense approval.
➁ The program includes advanced headgear with AR capabilities, and Anduril will manage production and development with Microsoft Azure as the preferred cloud provider.
➂ Palmer Luckey, the founder of Oculus VR and Anduril Industries, sees this as a deeply personal achievement, emphasizing the potential of the technology to save lives.
Hanwha Precision Machinery is on the verge of officially securing orders from SK Hynix for thermal compression (TC) bonders used in the production of high-bandwidth memory (HBM). Sources indicate that the equipment manufacturer is in the final stages of verification with the chip-maker.
➀ KLA报告第三季度营收31亿美元，非GAAP每股收益8.20美元；
➁ 预计未来季度营收30-35亿美元，非GAAP每股收益7.55-8.65美元；
➂ AI和HBM是推动高端市场增长的关键因素。
➀ Samsung has reportedly received approval to supply NVIDIA with its 8-layer HBM3E memory chips for less powerful AI GPUs destined for China.
➁ Despite previous distrust from NVIDIA's CEO, Samsung is now approved to supply these chips.
➂ SK hynix remains NVIDIA's main partner for advanced HBM memory, providing 12-layer HBM3E and working on next-gen HBM4 for future GPUs.
➀ NVIDIA is distancing itself from Samsung and its HBM memory due to trust issues, as stated by CEO Jensen Huang.
➁ Despite previous positive remarks, NVIDIA is working to certify Samsung's HBM3E memory for AI GPUs.
➂ If unsuccessful, NVIDIA may rely on SK hynix and Micron for HBM supply.
➀ SK hynix reported record profits in 2024, driven by high demand for its HBM memory used in AI chips.
➁ Despite the success, the labor union criticized the company for not honoring a profit-sharing agreement.
➂ The company plans to focus on sustainable growth and flexible investments.
➀ Tongfu Microelectronics has begun trial production of HBM2 memory, potentially for Huawei AI GPUs.
➁ This marks a significant step for China's domestic industry, despite being behind current HBM standards.
➂ HBM4 is on the horizon and will be featured in NVIDIA's next-gen Rubin R100 AI GPU.
