Qingrong New Material Technology (Jiaxing) Co., Ltd. has recently declared the successful closure of its angel round funding, amassing tens of millions of yuan. CAS Star spearheaded this funding round, with contributions from several other institutions. The proceeds are earmarked for scaling up production lines, advancing equipment R&D, and expanding market reach. Established in September 2024, the company is dedicated to the R&D and commercialization of high-performance functional composite film materials, with a primary focus on two key areas: high-energy-storage capacitor films and high-frequency copper-clad laminates. The company's technological foundation stems from a team at Tsinghua University, whose key members boast extensive expertise. Qingrong has overcome foreign technological barriers, achieving mass production of non-glass-fiber-cloth PTFE-based high-frequency copper-clad laminates. Its products are making inroads into high-end sectors, with its main offerings gaining certification from military enterprises and its high-energy-storage capacitor films forging partnerships with downstream clients. The novel composite dielectric film exhibits exceptional performance, rendering it ideal for high-end applications. CAS Star is of the opinion that Qingrong New Materials exhibits robust self-reliance and global competitiveness, with vast commercial potential.

On December 25, 2025, the China Securities Regulatory Commission (CSRC) approved the registration application for the Initial Public Offering (IPO) of GRINM Metal Composite Materials (Beijing) Co., Ltd. on the Science and Technology Innovation Board (STAR Market), with the company planning to raise 900 million yuan.

SK Hynix has moved up the mass production schedule of HBM4 at its newly established Cheongju M15X plant in South Korea—a facility specifically designated for HBM4 manufacturing—by four months, bringing the start date to February of the coming year. Initially, the plant will have a production capacity of roughly 10,000 wafers, with projections to scale up to tens of thousands of wafers by the year's end. The installation of pertinent equipment is proceeding swiftly.

This strategic adjustment is viewed as a preemptive move to meet NVIDIA's requirements, given that NVIDIA's next-gen AI accelerator, dubbed "Rubin," is anticipated to hit the market in the latter half of next year and will necessitate HBM4 memory. Presently, SK Hynix has furnished samples and is working in tandem with NVIDIA and TSMC for evaluation purposes.

The global HBM market is characterized by intense rivalry. Samsung has already supplied samples and carried out testing, whereas Micron's HBM4 production capacity for the upcoming year is completely booked. NVIDIA is slated to unveil its final supplier selection and share allocation early next year, with power consumption efficiency emerging as a pivotal assessment criterion. Samsung, for its part, is making concerted efforts to enhance its process yield.

On the afternoon of December 25th, market reports surfaced indicating that four prominent silicon wafer producers had jointly and substantially raised their prices. Specifically, the quoted price for 183N silicon wafers soared to 1.4 yuan per piece, 210RN wafers reached 1.5 yuan per piece, and 210N wafers climbed to 1.7 yuan per piece, marking an average price increase of 12%. Personnel from several silicon wafer manufacturers corroborated these reports. The industry consensus is that this price surge is primarily driven by the significant uptick in the costs of upstream silicon materials.

In November, NVIDIA's CEO, Jensen Huang, paid a visit to TSMC, where both parties engaged in in - depth discussions regarding the burgeoning demand for cutting - edge AI chips. This meeting has ignited a fresh surge of factory - building fervor at TSMC.

TSMC has promptly urged its upstream equipment suppliers to significantly shorten delivery cycles. This move has effectively catapulted the entire supply chain into what can be described as a 'wartime state.' It is anticipated that high - intensity shipments will persist uninterrupted until the second quarter of 2026.

In a bid to meet the soaring demand, TSMC is undertaking an extensive factory - construction spree in key locations such as Hsinchu and Kaohsiung. The construction projects encompass a wide range of initiatives, including the establishment of 2 - nanometer production lines, the expansion of existing 2 - nanometer production capacity, the acceleration of 3 - nanometer processes, the commencement of construction for a 1.4 - nanometer plant, and the expansion of capacity for 'advanced packaging' technologies.

Moreover, TSMC's business operations in the United States are also in full swing. Its first wafer fab in Arizona has already commenced mass production, and two additional fabs are currently under construction.

Industry experts have made the projection that TSMC's capital expenditures in the upcoming year will fall within the range of US$48 billion to US$50 billion. This substantial investment is aimed at bolstering its capacity expansion plans and ensuring that the semiconductor equipment supply chain operates at maximum capacity until 2026.

Based on the report from DigiTimes, because of the substantial price hike of DDR4, Samsung has decelerated the pace of phasing out DDR4 memory production. The company intends to ink long-term supply pacts with certain clients in the server industry during the first quarter of 2026. These contracts will feature 'non-cancellable and non-modifiable' terms to guarantee optimal profitability in capacity allocation.

On December 25, the schedule for the appointment (reserved for: "appointed") disclosure of 2025 annual reports for companies listed on the Shanghai Stock Exchange was announced. Pricewell Technology, a company listed on the Science and Technology Innovation Board, will be the first to disclose its annual report on February 3, 2026. On the main board, *ST Huawang will disclose its report on February 13, and Shangwei Co., Ltd. will disclose its report on February 14.

Tech media sources reveal that some tech-savvy netizens have conducted hands-on tests and discovered a notable performance boost in random read and write speeds after forcibly activating native NVMe driver support in Windows 11. This improvement is particularly pronounced in 4K random read and write operations. In contrast, the enhancement in sequential read and write speeds is relatively modest. NVMe SSDs have firmly established themselves as the market standard; however, both Windows 10 and Windows 11 systems still interact with them via a SCSI emulation layer. While this setup guarantees compatibility, it also places an additional burden on the CPU and introduces higher latency, thereby constraining the NVMe SSDs' theoretical performance potential. Microsoft has addressed this issue by introducing a native NVMe driver in Windows Server 2025, which eliminates the need for the SCSI emulation layer. The company touts an impressive IOPS performance increase of up to 80%. Although a handful of users have managed to enable the native NVMe driver on Windows 11 by tweaking the registry and have reported performance enhancements, the feature remains unstable at present and carries risks such as the potential for hard drives to become inaccessible. Consequently, it is not advisable for everyday users to activate this feature in their regular computing routines.

In December 2025, the prices of DDR5 memory soared dramatically. This surge was driven by a surge in AI demand and adjustments in manufacturers' production capacities. Some DDR5 memory models even became more expensive than graphics cards. Against this backdrop, the Russian hardware outlet PRO HiTech put forward a radical solution. Drawing on the modification expertise of hardware expert Viktor "Vik-on", hardcore DIY enthusiasts are now attempting to hand-assemble memory modules in an effort to sidestep the exorbitant costs.

This phenomenon underscores profound shifts within the DIY market ecosystem. Since early 2025, when DDR4 memory faced structural shortages due to production capacity relocation, consumer memory prices have surged by over 200%. This price hike has directly led to a 40%-50% year-on-year decline in motherboard shipments. In response, player strategies have become increasingly polarized. Some users have resorted to temporary cost-saving measures, such as using single-stick memory configurations or laptop RAM adapters. Others have chosen to upgrade older platforms, like extending the lifespan of AMD AM4 motherboards by simply swapping out CPUs.

Market fragmentation is also intensifying. Gaming consoles and mobile devices are luring PC users away with their cost-effectiveness. Meanwhile, high-end DIY enthusiasts are focusing on personalized customization, pursuing premium upgrades like chassis liquid cooling systems. Industry forecasts indicate that AI-driven memory shortages may persist until 2026. If prices remain high over the long term, the DIY market will likely accelerate its shift towards niche specialization. Traditional value-oriented builds will face a complete restructuring.

On December 25th, news emerged stating that the release timetable for the 2025 annual reports of companies listed on the Shanghai Stock Exchange has been made public. PrimeNano, a company traded on the Science and Technology Innovation Board (STAR Market, a board under the Shanghai Stock Exchange focusing on innovative and high-tech enterprises), is set to take the lead by unveiling its annual report on February 3, 2026. Following closely behind, main board companies *ST Huawang and Shangwei Share will present their respective reports on February 13 and February 14, 2026.

Recently, the x86 chip "Serpent Lake," a collaborative effort between Intel and NVIDIA, has made its debut, directly challenging AMD's high-performance APU offerings, including the Strix Halo. According to RedGamingTech, Serpent Lake is far from a mere packaging solution. The CPU component is anticipated to leverage Intel's Titan Lake architecture, a successor to Nova Lake. Meanwhile, the GPU segment will seamlessly integrate NVIDIA's next-generation Rubin architecture, following in the footsteps of Blackwell.

Furthermore, Serpent Lake is projected to incorporate 16x LPDDR6 high-bandwidth memory and may adopt a design where the memory is closely packaged alongside the processor, a move aimed at enhancing overall performance. From a manufacturing perspective, the product is expected to be crafted using TSMC's advanced N3P process technology.

Serpent Lake is primarily designed for high-end mini PCs, high-performance laptops, and handheld gaming consoles that require exceptional graphics capabilities but are limited by size constraints, preventing the integration of discrete graphics cards. While the Serpent Lake project has yet to be officially confirmed by either Intel or NVIDIA, NVIDIA's previous substantial investments and the technical frameworks outlined by both companies suggest that the x86 platform is on the verge of receiving a significant boost from NVIDIA's cutting-edge graphics technology.

Should this plan materialize, it will pose a direct threat to AMD's Strix Halo and upcoming Medusa Halo products, intensifying the competition in the high-performance computing market.

Rumors had been circulating that NVIDIA was set to acquire AI chipmaker Groq in a deal valued at around $20 billion. However, NVIDIA has issued a response, clarifying that it has not acquired Groq. Instead, the tech giant has secured a non-exclusive license to utilize Groq's intellectual property and has recruited top-notch engineering talent from Groq's ranks. The collaboration between the two entities aims to jointly deliver cutting-edge accelerated computing technology. Notably, Groq will maintain its status as an independent entity, continuing its operations without being subsumed into NVIDIA.

On the 25th, He Yongqian, the spokesperson for the Ministry of Commerce, addressed the US's announcement regarding the imposition of additional 301 tariffs on select Chinese semiconductor products. He stated that China has duly noted the situation and, through the China-US economic and trade consultation mechanism, has formally lodged representations with the US. China disagrees with the findings of the US 301 investigation and stands firmly against the US's move to levy extra tariffs on Chinese semiconductor products.

He Yongqian highlighted that the US's unilateral tariff actions breach World Trade Organization regulations, disrupt the global economic and trade landscape, cause instability in the worldwide industrial and supply chains, and adversely affect the interests of both US businesses and consumers. China urges the US to promptly rectify its errors and revoke the relevant measures. China is open to collaborating with the US, adhering to the principles of mutual respect, peaceful coexistence, and mutually beneficial cooperation, to address concerns through equitable dialogue and consultation. Should the US persist in infringing upon China's rights and interests, China will not hesitate to take the necessary steps to robustly defend its own rights and interests.

Multiple reliable sources have reported that NVIDIA has communicated to its Chinese clientele its intention to supply the H200, its second-most powerful AI chip, prior to the Lunar New Year, which falls in mid-February 2026. The anticipated total shipment quantity ranges from 5,000 to 10,000 chip modules, translating to roughly 40,000 to 80,000 individual chips. In response, NVIDIA clarified that it is consistently overseeing its supply chain operations and assured that the provision of H200 chips to its authorized Chinese customers will not compromise its global supply capacity.

The H200 chip boasts 141GB of HBM3e memory along with a staggering 4.8TB/s bandwidth, marking a substantial leap in performance compared to its predecessors. This enhancement adequately caters to the computational requirements of the vast majority of domestic commercial applications. Nevertheless, it is important to note that the H200 chip has not yet secured the definitive import approval from Chinese regulatory bodies. Consequently, the exact timeline for delivery remains subject to uncertainty.

According to the announcement on the China Government Procurement Network, Shanghai Micro Electronics Equipment (Group) Co., Ltd. has won the bid for the project of purchasing a step-and-scan lithography machine (procurement by zycgr22011903), with a quantity of one unit and a transaction value of approximately 110 million yuan. The model of the winning bid goods is SSC800/10.

In December 2025, the anticipated domestic polysilicon output is set to hit 113,500 metric tons. This figure indicates a minor surplus when compared to the demand for wafers. Looking ahead to January 2026, it's projected that production will remain above 100,000 metric tons. Despite the fact that certain manufacturers have formulated plans to scale back production, the magnitude of this reduction will be relatively constrained.

Recently, there has been an uptick in cell prices. In response, manufacturers have taken the initiative to regulate their shipping speeds as a means of sustaining these price levels. Among downstream buyers, with the exception of those with urgent and indispensable demand, the majority are adopting a wait-and-watch approach. They are exercising caution when making purchases. The intensifying tug-of-war between upstream and downstream players has resulted in an accumulation of cell inventories. Specifically, the inventory levels in the upstream segment of the photovoltaic cell market have continued to edge up on a week-by-week basis.

On the afternoon of December 25th, the photolithography machine concept witnessed a sudden and significant upward spike. Guofeng New Materials soared straight to the daily limit, marking its second such achievement within a mere three - day span. Additionally, multiple stocks, such as Chinasoft Technology and Huarong Chemical, also experienced substantial surges.

On December 25th, the General Office of the Shanghai Municipal People's Government unveiled its strategic blueprint, [specific document name if available, otherwise omit], outlining a vision to foster emerging industry clusters. Central to this initiative is the establishment of a national local hub dedicated to 6G comprehensive trials. The plan encompasses a multifaceted approach to propel 6G technology forward, including intensive research and development, formulation of industry standards, rigorous technical trials, scalable pilot production, innovative scenario-based applications, and robust international collaboration. In a show of commitment, the government pledges to offer staged financial support, covering up to 30% of the total project investment, with a cap set at 30 million yuan, adhering strictly to regulatory guidelines.

Industry data on Thursday revealed that Samsung Electronics has emerged as the most gifted stock among South Korean retail investors this year. According to an analysis by KB Securities, through its stock gifting service, Samsung Electronics topped the list in terms of transaction volume, followed by shares of LG CNS, a subsidiary of LG Electronics, and preferred shares of Samsung Electronics.

On December 24, NavInfo released an announcement stating that it would postpone the timeline for certain investment projects funded by raised capital and make adjustments to some sub-projects. Specifically, the originally planned operational dates for the 'Smart Connected Vehicle Chip R&D Project' and the 'Autonomous Driving Exclusive Cloud Platform Project' have been pushed back to December 2028 and December 2026, respectively. In addition, the company intends to discontinue two sub-projects under the 'Smart Connected Vehicle Chip R&D Project' — the Low-cost DA Chip AC8005 and the Visual Processing Chip AC6815 — and introduce two new ones: the Cost-effective Connected Instrument SoC AC8117 and the High-performance 5G Connected Car IVI SoC AC8277. It is worth noting that these adjustments will not affect the total amount of raised funds.