chips

Analyst jukan from Citrini posted on platform X that, according to WSJ, Apple has obtained a semiconductor tariff exemption after committing to use Intel's foundries. Apple will not only produce Mac chips at Intel but will also produce iPhone chips.

According to a report by The Information, multiple informed sources revealed that China's leading AI company Zhiyu is considering designing its own AI chips. It is reported that as the demand for its GLM series large models has rapidly surged, coupled with increasing constraints on computing resources, Zhiyu has recently had preliminary contacts with several domestic chip design companies to explore the possibility of collaborating on the development of custom AI processors.The report pointed out that for large AI model developers, the layout of self-developed chips aims to achieve software and hardware collaborative optimization, improve computing power utilization efficiency, and reduce dependence on external GPU suppliers. In addition to Zhiyu, another domestic large model company, DeepSeek, has also been reported to be simultaneously advancing its self-developed chip project focused on inference scenarios. Globally, top AI developers such as OpenAI (which has announced its self-developed chip Jalapeño) and Anthropic are actively laying out custom chips, and the AI industry is accelerating towards a dual-line layout phase of "model + chip" software and hardware integration.

According to the latest report from Omdia titled "2026 Q2 Semiconductor Application Market Forecast Tool (AMFT)", driven by the large-scale development of AI infrastructure, the expected year-on-year growth rate of China's semiconductor market size in 2026 has been significantly revised upward to 92.9%, reaching a total of $812.08 billion. Among them, the computing and storage category is expected to grow by 126%, accounting for 62.9% of the overall application market, marking that the industry is fully entering the AI-driven era in sync with the global trend.In the segmented chip categories, influenced by the surge in demand for AI cloud and edge inference leading to tight supply and demand, the market size of China's storage chip market is expected to skyrocket by 262.9% to $449.6 billion, with market share jumping to 55.4%. At the same time, AI demand has also driven significant growth in logic chips (27.9%), analog ICs (25.4%), and microcontrollers (15%), creating enormous space for the improvement of local semiconductor self-sufficiency and capacity utilization. In contrast, the wireless communication category, primarily focused on smartphones, although experiencing an absolute scale growth of 68.8%, is expected to see its overall market share decline from 30.43% in 2025 to 26.63% due to the dual pressure of soaring storage costs and shrinking shipment volumes.

According to the Goldman Sachs report on "China's AI Computing Power" disclosed by P Equity Research, China is accelerating the construction of a national computing power network, with related infrastructure projects expected to attract 7 trillion yuan in investment by 2026. In the next five years, investment in data centers is projected to reach about 2 trillion yuan. Currently, funds and technology are being massively transferred to western computing power hubs, while data centers in first-tier cities are transforming to focus on ultra-low latency computing, edge nodes, and AI inference. Although GW-level clusters with over 100,000 chips remain scarce domestically, in typical GW-level computing power parks, workloads are primarily composed of inference, which accounts for more than half, as well as training and full-stack R&D.The report predicts that by 2026, the market share of domestically produced AI acceleration chips is expected to exceed 50%. Among them, Huawei and Alibaba's Pingtouge lead the domestic camp with shares of 20% and 7%, respectively, but Nvidia currently maintains overall market dominance with a 55% share. In terms of cost and performance, domestic chips have capital expenditures on IT power consumption that are 40% to 50% lower than imported chips, but due to performance gaps, their capital expenditures per unit of computing power are 2 to 4 times that of imported chips, and the computing power generated per unit of power consumption is only 10% to 30% of that of imported chips. Additionally, the daily average token output of Huawei's 910B/910C servers is about one-sixth to one-third of that of Nvidia's H800, resulting in significantly lower API profit margins based on that hardware compared to peers using Nvidia hardware.

The South Korean government has announced significant project plans in the areas of chips, physical artificial intelligence, and artificial intelligence data centers. South Korean President Lee Jae-myung stated that it is necessary to ensure the development of artificial intelligence elements faster than other countries, and chip production construction must be completed as soon as possible.The current factory site selection is nearing its limits in terms of water resources and other infrastructure. The southwestern region will invest between 50 trillion to 200 trillion won in the new projects, while Gwangju and Jeolla may invest 52 trillion won in the projects. Regarding the detailed planning for the factories, the South Korean government announced that it is expected to build four chip factories in the southwest, with an investment of about 800 trillion won; Samsung Electronics and SK Hynix will each build two new factories; it is expected that investment in the chip sector will reach at least 300 trillion won over the next 15 years, including next-generation memory, edge artificial intelligence, and defense.Investment in the chip packaging cluster in the Chungcheong region is expected to reach 81 trillion won. It is anticipated that approximately 55 trillion won will be invested in the construction of artificial intelligence data centers. The South Korean government expects to double DRAM production capacity within five years, and the global memory market is expected to quadruple in five years. The South Korean side also stated that it seeks to distribute the growth dividends to the public.

According to Yicai, there were previous reports that Kunlun Core, a subsidiary of Baidu, plans to go public in Hong Kong, and Tencent has become its client, while ByteDance is also considering using its AI chips. In response to this, sources close to the matter told Yicai today that ByteDance currently has no intention of cooperating with Kunlun Core.

According to the Financial Times, Apple is lobbying the Trump administration for approval to purchase memory chips from Chinese chipmaker Changxin Storage. Due to multiple factors leading to a rise in memory prices, Apple unusually raised the prices of MacBook and iPad on Thursday, resulting in a market value loss of $263 billion. Apple hopes to alleviate the financial pressure from suppliers by including Changxin Storage in its list of memory suppliers.Changxin Storage and Yangtze Memory Technologies have currently been placed on the U.S. Department of Defense's blacklist of Chinese military companies (the 1260H list). The U.S. Department of Commerce had planned to add Changxin Storage to the entity list last year, but the White House postponed the plan due to ongoing trade negotiations with China at that time. Currently, aside from Chinese companies, the global memory chip industry is mainly dominated by three companies: U.S.-based Micron, and South Korea's Samsung and SK Hynix.

According to Jinshi reports, Broadcom is expected to deploy OpenAI chips faster than anticipated.

According to on-chain analyst Yu Jin's monitoring, within 2 days, the SIREN controllers have essentially "broken down" 680 million SIREN (accounting for 94% of the total). This sell-off of SIREN led to a price drop of 96% (from $1.3 to $0.05), resulting in approximately 64.8 million USDT. About 200 million of these flowed into centralized exchanges such as Binance, Gate, and KuCoin, with most being bought in small amounts on-chain by hundreds of addresses after the price fell below $0.1. These small purchases are likely still orchestrated by the controllers, aiming to recover chips at low prices, increase tracking difficulty, and prepare for the next manipulation.

Recently, the global memory chip sector has been active, with the expansion of AI computing power and the ongoing construction of data centers continuously driving market demand for high-performance storage products, attracting widespread attention from investors. According to Gate market data, Micron Technology (MU) reached a high of $933.49 in 24 hours, currently reported at $906.52, with a 24-hour increase of 3.49%; SK Hynix (SKHYNIX) reached a high of $1348.10 in 24 hours, currently reported at $1318.40, with a 24-hour increase of 8.31%. According to CoinGlass data, the trading volume of Gate MU (Micron Technology) contracts reached $11.3353 million in 24 hours, and the trading volume of Gate SKHYNIX (SK Hynix) contracts reached $3.9042 million in 24 hours, both ranking among the industry leaders, with the trading heat of related assets continuing to rise.

According to the Financial Times, Intel (INTC.O) plans to launch an artificial intelligence chip by the end of the year, which will use cheaper memory and cooling technologies than competitors from Nvidia and AMD.

Glassnode disclosed the latest on-chain data showing that the Bitcoin spot trading volume difference (7-day moving average) has re-entered a sustained negative range. As the Bitcoin price fell back to around $73,000, the market experienced significant selling pressure accelerating, with sellers continuing to dominate the spot market, and the pressure of chip distribution is gradually accumulating.Recently, during the process of Bitcoin falling back from above $80,000 to around $75,000, spot demand, ETF fund inflows, and volatility expectations have all continued to weaken. Although market positions have been somewhat reset, investor confidence remains limited, presenting an overall cautious wait-and-see attitude.

Some high-end graphics cards, servers, and other products are gradually landing in JD's self-operated channels. According to JD's data, since May 14, the attention on "NVIDIA" H200 related chips and server workstation equipment launched on JD's platform has significantly increased, with the search volume for NVIDIA H200 related chips increasing 25 times year-on-year, and the search volume for NVIDIA H200 related servers and workstations increasing over 40 times year-on-year.

According to a report by Reuters, the United States has approved the sale of Nvidia H200 AI chips to about 10 Chinese companies, including Alibaba, Tencent, and ByteDance, but no chips have been delivered yet, and the related transactions are at a standstill.The report pointed out that Nvidia CEO Jensen Huang joined the delegation during President Trump's visit to China, trying to promote the sale of H200 chips in China. The U.S. Department of Commerce has also approved some distributors, such as Lenovo and Foxconn, to directly purchase and resell the chips. However, due to strict security, usage, revenue sharing, and legal restrictions imposed by the U.S. on H200 exports, there are conflicts with China's supply chain security policies, leading to stricter order approvals.