nvidia

NVIDIA (NVDA.O) and Oracle (ORCL.N) are expanding the deployment scale of the NVIDIA Vera Rubin platform.

The semiconductor research organization SemiAnalysis has released the actual test results of Huawei's Ascend 950PR. The organization selected 1,055 sets of commonly used matrix calculations for large models. Huawei's own underlying program has allowed the 950PR to achieve 86% of its theoretical computing power, higher than NVIDIA's H100 at 75.5%; however, the actual computation speed is still only 43% of the H100, and only 21% compared to the relatively newer B300. This set of results has not yet utilized the newly ported operators from DeepSeek.DeepSeek is working on reducing data movement and waiting from the software side. In attention computation, after adjusting storage with FlashMLA, the two sets of data that originally needed to be read separately can now be retrieved at once; TileLang allows the chip to compute while preloading the next batch of data. SemiAnalysis measured that automatic scheduling has reduced the time for some small computing tasks by up to about 40%. DeepSeek has also ported DeepEP, which is responsible for MoE expert communication, but communication and computation will still compete for memory bandwidth, and Huawei plans to continue improving the firmware.Last year, DeepSeek attempted to train R2 using Huawei's Ascend chips, but even with Huawei engineers on-site for assistance, it could not be completed due to issues related to chip stability, inter-card communication, and software support. This year, both parties continue to collaborate, and by the end of September, DeepSeek focused on publicly releasing six sets of Ascend adaptation code. In October, it was reported that Huawei is adjusting the next generation of chip designs according to DeepSeek's model requirements. The 950PR tested this time is mainly aimed at inference input processing, while DeepSeek is preparing to use the 950DT for large-scale pre-training as early as the end of the year. Previously, both parties had conducted tests at the thousand-card level, but to enable tens of thousands of chips to complete a round of large model training continuously, more complex coordination and stability issues still need to be resolved.

According to Gate data, the "Seven Sisters" sector in the U.S. stock market is generally down before the market opens, with mixed performance in the storage sector.NVIDIA is down 0.53%, Apple is down 0.27%, Microsoft is down 0.59%, Google is up 0.16%, Amazon is down 0.10%, Meta is down 0.36%, and Tesla is down 0.21%.Micron is up 0.55%, SanDisk is up 0.65%, SK Hynix is up 1.67%, Western Digital is down 0.58%, and Seagate Technology is down 0.34%.

According to Jinshi News, Nvidia will invest 3 billion dollars in equity issuance to Meta, and both parties will also expand their cooperation in AI infrastructure and models.Both sides will consider collaboration around the OpenShell reference architecture and Nvidia Sentry, and will expand the deployment scale of Nvidia's products such as Vera Rubin, Vera CPU, Spectrum-X, and BlueField-4, while further deepening cooperation in MUSE and the MUSE series of models.

NVIDIA (NVDA.O) and Meta will expand the deployment scale of NVIDIA Vera Rubin, Vera CPU, Spectrum-X, and BlueField-4.

NVIDIA (NVDA.O) and Meta are considering collaboration on the OpenShell reference architecture and NVIDIA Sentry.

NVIDIA (NVDA.O) announced an investment of 3 billion dollars in equity issuance to Meta.

Market news indicates that Nvidia and Meta will expand their collaboration on MUSE and the MUSE series models against a backdrop of steady growth.

IBM (IBM.N) and NVIDIA (NVDA.O) will bind AI entity identity with runtime access control.

According to Goldman Sachs data, in the past two weeks, global technology stocks have experienced a net sell-off of approximately $20 billion, with the intensity of the sell-off reaching an extreme level of -3.3 standard deviations. The sell-off scale in the TMT sector is close to the $23 billion seen in July of this year. Meanwhile, the volatility spread between the Nasdaq 100 Index and the S&P 500 Index has risen to the 98th percentile of the past five years, indicating a significant increase in market demand for hedging against technology stocks.According to JPMorgan, the scale of debt financing for AI infrastructure has exceeded $500 billion this year, with the financing costs for large-scale cloud service providers rising by about 120 basis points compared to before. As Google, Microsoft, and Amazon release their earnings reports this week, the market will focus on whether AI capital expenditures can be converted into actual revenue, as well as the sustainability of financing for subsequent investment expansions.JPMorgan believes that the financing relationship between Nvidia and GPU operators may become a channel for credit risk transmission. If operator financing tightens, pressure may spread upstream along the supply chain. The bank maintains a constructive view on the fundamentals of the technology sector, believing that a buying window may appear in early November, but it is still necessary to observe signals such as further market position clearing, corporate earnings realization, and credit spread narrowing.

Market news reports that Nvidia, AMD, and Broadcom intend to purchase up to 5 billion dollars of Meta Platforms stock.

Samsung Electronics and SK Hynix plan to start mass production and delivery of 8-layer HBM4 in the second quarter of next year, supplying it to NVIDIA.According to informed sources, NVIDIA will equip the 8-layer HBM4 based on the scale of the Rubin Ultra NVL rack, and it is expected that the proportion of 8-layer HBM4 will significantly increase starting from the middle of next year.

According to a report by Reuters cited by Phoenix Technology, on October 26, Beijing time, Elon Musk, the founder of SpaceX, stated that the company has completed $40 billion in debt financing for the procurement of graphics processors. Musk posted on the social platform X that this financing is primarily for the purchase of GPUs that will be delivered in the first and second quarters of next year, and the debt can be redeemed early, which will not have a significant impact on this quarter.Musk added that SpaceX's investment-grade credit rating, along with the revenue generated from Starlink satellite internet and launch services, has helped the company maintain its ability to secure credit financing. Earlier this month, Reuters reported that SpaceX is in talks with banks and asset management companies to raise $40 billion for the purchase of NVIDIA's AI chips.For the three months ending September 30, SpaceX's revenue grew by 161% year-on-year, reaching $14.58 billion. During the same period, the company's capital expenditures on computing infrastructure nearly doubled, reaching $19.6 billion.

Dongfang Gangwan Fund Chairman Dan Bin is making a "major adjustment" in the technology sector. According to statistics from Private Equity Ranking Network, as of the end of the third quarter of 2026, Dongfang Gangwan's overseas fund held a total of 13 U.S. stock targets, with a total market value of approximately $1.469 billion.From the perspective of industry distribution, AI remains a key focus for Dongfang Gangwan. As of the end of the third quarter, Google was the largest holding in Dongfang Gangwan's overseas fund. However, in terms of buying and selling actions, Dan Bin's preference for sub-sectors within the AI industry has changed significantly compared to the second quarter.According to statistics, in the third quarter of this year, Dongfang Gangwan's overseas fund reduced its holdings in targets such as Advanced Micro Devices (AMD), Broadcom, and Lumentum, which were newly purchased in the second quarter, while increasing its positions in storage giant Micron Technology, SanDisk, as well as Nvidia, Google-C, and Taiwan Semiconductor Manufacturing Company.

Rumors of the discontinuation of certain models of NVIDIA consumer graphics cards have stirred the Huaqiangbei market, with graphics card prices changing daily. The official initial retail price of the RTX 5090 is approximately 15,000 to 16,000 yuan, while current channel prices have soared to 55,000 to 60,000 yuan.The initial price of the RTX 5080 was about 8,000 yuan, and it has now risen to 17,000 yuan.

AI video startup Higgsfield Inc. announced that it has received investments from NVIDIA and TCV. This round of financing will be used to support global infrastructure development and talent recruitment. The company is headquartered in San Francisco, with its engineering team primarily located in Kazakhstan.This financing is a follow-up extension to the Series B funding in August, during which it raised $400 million, bringing the company's valuation to $5.4 billion.

Korean storage substrate manufacturer TLB has started the development of the AI server low-power memory module specification SOCAMM3 substrate with two large domestic and foreign storage manufacturers. According to industry news on the 22nd, TLB recently began negotiating design directions and manufacturing processes with two existing storage customers, with specifications yet to be determined. The storage manufacturers plan to complete the design with the substrate manufacturers before finalizing the specifications, with actual substrate mass production expected in 2028.SOCAMM will make the low-power DRAM for mobile devices into a detachable AI server module, which will be first used by NVIDIA in its own AI servers. SOCAMM2 has developed into a JEDEC standard. Currently, each NVIDIA Vera CPU is equipped with 8 SOCAMM2 based on LPDDR5X, with a maximum capacity of 1.5TB per CPU and a bandwidth of 1.2TB per second. The SOCAMM2 substrate requires five layers, and the number of layers for SOCAMM3 will increase, raising the difficulty of processes such as micro-hole processing. The applicable systems have not been disclosed, and the mass production timeline in 2028 coincides with NVIDIA's roadmap. NVIDIA announced in March this year at GTC that it will launch the Feynman GPU and the new CPU Rosa in 2028.TLB's SOCAMM2 substrate has already entered mass production earlier this year, with production volume and sales in the second quarter increasing more than three times compared to the previous quarter. The production time and unit price are both about twice that of existing module substrates. The industry expects TLB's SOCAMM substrate sales to be around 70 billion Korean won this year, and about 140 billion Korean won next year. Brokers estimate the company's overall revenue this year to be about 350 billion to 370 billion Korean won, a year-on-year increase of about 40%, and around 450 billion Korean won next year. The company will begin construction of its second factory next month in the An Phu Industrial Park in Bac Ninh Province, Vietnam, aiming for production in the fourth quarter of next year, but SOCAMM substrates will still only be produced at the headquarters in Ansan in the short term.

According to the Science and Technology Innovation Board Daily, Nvidia CEO Jensen Huang stated during a media interview at the GTC conference in Berlin on Thursday that the likelihood of a memory supply surplus by 2029 is low. He mentioned that one great aspect of memory is that there is always more that can be used, and he noted that HBM will continue to be applied in the coming years.Jensen Huang also stated that Nvidia is the first company in the world to introduce LPDDR memory, used in mobile phones, into data centers, which doubles its energy efficiency compared to any current global solution, thus it will continue to use this type of memory extensively. He mentioned last Thursday that customer demand for AI chip orders continues to exceed the company's supply capacity, with new orders pouring in almost weekly, if not daily.In the past week, storage chip stocks faced sell-offs due to market concerns over supply surplus and weak demand. A well-known media outlet in South Korea published a column last week, citing two experts who predicted that by 2029, there will be a significant increase in semiconductor supply, and the storage chip market may turn into a red ocean. On Thursday, the Nasdaq Composite Index closed down 2.6%, at 26008.86 points; SanDisk fell 8.51%, SK Hynix fell 8.07%, Seagate Technology fell 7.86%, Western Digital fell 6.43%, and Micron Technology fell 6.23%.