SK Hynix will now surely adjust its DRAM production capacity projection as it now suppliers HBM to both Nvidia and Broadcom. It is currently supplies the majority of HBM used by Nvidia for their AI accelerators.

Samsung Electronics is expected to supply its cutting-edge chips to the top graphics processing unit maker Nvidia soon, stepping closer to securing a footing in the lucrative AI Chip market.

The HBM4 chip could also be used in Tesla’s AI data centers under development and its self-driving cars, which are currently fitted with HBM2E chips for pilot programs.

According to informed sources, Samsung is currently discussing plans to redesign some of the 1a DRAM circuits internally, but has not yet made a final decision as this decision involves various risks.

TSMC has revealed its upcoming plans in the advanced packaging technology sector, emphasizing the significance of integrating AI Chip memory and logic chips through 3D IC technology. The company anticipates a substantial reduction in production costs for AI processors and aims to maintain a leading position in packaging innovation by 2027.

SK Hynix, a key player in the semiconductor industry, is set to enhance its High Bandwidth Memory (HBM) offerings by introducing customizable base die technology. This move is aimed at providing clients with tailored solutions that can improve chip efficiency and performance. The company also plans to incorporate chiplet technology into SSD controllers, marking a significant advancement in storage device design.

Taiwan Semiconductor Manufacturing Company (TSMC) is gearing up for the mass production of its cutting-edge A16 process technology, slated for the second half of 2026. This advanced node has already attracted its first customers, including tech giant Apple and artificial intelligence (AI) powerhouse OpenAI, who have placed orders for their AI Chips.

South Korean AI Chip design company Rebellions is on track to unveil its next-generation AI Neural Processing Unit (NPU) chip, REBEL, by the end of 2024. The chip is designed to accelerate large language and multimodal models and will feature Samsung's 4nm process technology and HBM3E 12H memory, along with support for 800Gb Ethernet.

The latest report from SEMI indicates a continued positive trend in the global semiconductor industry for Q2 2024, with significant growth in integrated circuit sales and stable capital expenditures. Despite a slowdown in recovery in some end markets affecting H1 growth, the industry is experiencing a strong tailwind from surging demand for AI Chips and High Bandwidth Memory (HBM).

Korean artificial intelligence (AI) processor manufacturers Rebellions and Sapeon Korea are merging, potentially positioning them as strong competitors against global industry leaders like Nvidia. The merger, announced by SK Telecom, comes with an estimated value of 1.14 trillion won ($851.2 million) and is expected to bolster the companies' market presence. Rebellions, which is also preparing for an IPO, is valued between 2 to 4 trillion won by industry experts. The merged entity is backed by significant global financial investors and Korean telecom giants, aiming to supply products to global data centers within the year.

Despite a sluggish market for EVs and smartphones, Japanese semiconductor equipment manufacturers have reported an 80% increase in operating profits for the second quarter, thanks to robust demand from the AI sector.

The international semiconductor industry organization, SEMI, has emphasized the need for more international standards in the backend or post-production processes of the chip industry. According to SEMI's Japan office President, Jim Hamajima, the lack of standardization is affecting the efficiency and profitability of companies like Intel and TSMC. He suggests that unifying standards for automation and materials specifications could streamline the work of equipment and material suppliers, allowing the industry to compete more effectively in meaningful areas.

SK Hynix is set to employ TSMC's N5 process foundation die for the construction of its HBM4 memory, marking a significant advancement in high-performance memory solutions. With the JEDEC standards for HBM4 nearing finalization, the industry anticipates the launch of SK Hynix's inaugural 12-layer stacked HBM4 products in the second half of 2025. This collaboration between TSMC and SK Hynix, solidified by a memorandum of understanding signed in April, is expected to propel the high-performance computing and AI Chip ecosystem to new heights with enhanced memory bandwidth.

Samsung Electronics is actively expanding into the artificial intelligence (AI) memory market by developing customized high bandwidth memory (HBM). The company revealed at the "Samsung Foundry Forum 2024" on July 9th that it is working with key clients such as AMD and Apple on customized collaborations. Samsung anticipates that the commercialization of its custom HBM will coincide with the mass production of HBM4, offering enhanced performance, power, and area (PPA) options that provide greater value compared to existing products.

Samsung Electronics Co., has officially launched a team dedicated to developing advanced high-bandwidth (HBM) memory, a core chip to power artificial intelligence (AI) devices.