meter

According to The Bell, Samsung Electronics' foundry division has delayed its contract for application processor foundry production with Qualcomm due to differences in production unit prices, and both parties have failed to narrow the price gap, prolonging negotiations. There have been no significant technical issues, and the currently discussed 2-nanometer process has received a considerable degree of stabilization evaluation.Industry sources indicate that the Qualcomm AP foundry using Samsung's 2-nanometer process may be postponed until next year, and it is unclear whether production can commence within this year. Qualcomm had previously entrusted advanced AP production to Samsung before 2021, and Samsung has been continuously striving to regain this business in recent years, having secured clients such as Tesla and Broadcom.Informed sources state that the contract delay has made it difficult for Samsung to produce APs for Qualcomm this year, and both parties may turn to discuss the next generation of products. After acquiring strategic clients, Samsung is no longer willing to accept low-price orders as it did in the past.

The Central Science Management Bureau confirmed on the 9th that TSMC's Central Science Phase II 1.4 nanometer factory expansion is fully accelerating. The first P1 factory has completed its steel structure and is expected to begin trial production in April next year, with mass production anticipated in the second half of next year, ahead of the originally scheduled mass production in 2028. TSMC has applied to the Central Science Management Bureau to set up two temporary offices at the site, which are expected to be completed in April next year, with the first batch of over 5,400 operational and outsourced personnel moving in.The advanced process new factory for TSMC's Central Science Phase II park broke ground last October, planning to build four 1.4 nanometer factories, with nearly 2,000 workers working day and night. The P1 factory is currently undergoing floor and exterior wall construction, with the factory building expected to be completed early next year. The P2 factory has begun basic construction and is scheduled to be completed in October next year, with both factories expected to start mass production successively next year. The P3 factory has obtained a construction permit, while the P4 factory is in the process of applying for a construction permit, planning to be built with a six-month gap. P3 is expected to be completed in the second quarter of 2028, and P4 is scheduled for completion in the fourth quarter of the same year. After the P2 factory is completed in the second half of next year, an additional 1,000 operational personnel will be added, with the total number of employees expected to be between 9,000 and 10,000 when all four new factories in Phase II are completed and put into production.

According to South Korea's "Cultural Daily," Samsung Electronics' wafer fab in Taylor, Texas, is expected to enter trial operation around the end of this month, as orders for 2-nanometer chips from Tesla, Broadcom, Arm, and others continue to pour in, with pre-production capacity nearing full load. The plant has accumulated an investment of $37 billion since 2022, with plans to begin trial operation early next month and mass production next year, with an initial monthly capacity of about 50,000 wafers, close to TSMC's approximately 80,000 wafers.The report states that the Taylor plant will produce Tesla's next-generation artificial intelligence chip AI5, Broadcom's new generation communication chips, and Arm's smart device AI chips, among others. Samsung's 2-nanometer yield has increased from about 60% earlier this year to around 80% recently. The company sent personnel for process, equipment, and yield to the plant in June, and began installing equipment in July. At a recent earnings meeting, Samsung projected that this year's 2-nanometer order volume would more than double compared to last year, and stated that NVIDIA's inference chip Grok3 began mass production last month; Samsung also recently raised the foundry price for 4-nanometer chips by up to 15%.The second Taylor plant is scheduled to start construction by the end of this year and begin mass production in 2030. Samsung has requested major partners to complete the relevant production equipment safety certifications in advance, and some partners are considering sending dozens of personnel to the local area. At the same time, Intel issued about $20 billion in new shares last month and expanded investments in 2-nanometer level 18A and 1.4-nanometer level 14A, while Japan's Rapidus completed its 2-nanometer factory in Hokkaido last month and entered the wafer trial production stage.

Samsung Electronics' wafer foundry division has recently surpassed 50% in the allocation of production capacity for the sixth-generation high bandwidth memory HBM4 substrate chips in the 4-nanometer process. The industry states that this division will invest more than half of its 4-nanometer wafers into HBM4 substrate chip manufacturing in the second half of this year to expand supply to companies like NVIDIA, Broadcom, and AMD.Samsung initiated mass production and shipment of HBM4 in February this year and plans to increase supply starting from the third quarter, significantly increasing the related wafer input since mid-year. This chip is based on Samsung's wafer foundry 4-nanometer (SF4) process. As of last month, over 50% of its total 4-nanometer capacity has been allocated to HBM4 substrate chips. According to insiders, Samsung's 4-nanometer process is currently at full capacity, with HBM4 substrate chips accounting for about 50% to 60%, and this high proportion will be maintained in the second half of the year.Samsung's wafer foundry production lines at Pyeongtaek Plant 2 and Plant 3 are mass-producing processes ranging from 4 to 7 nanometers, with a monthly capacity of about 30,000 wafers for the 4-nanometer process. Based on this, the wafer input related to HBM4 substrate chips is estimated to be about 15,000 wafers per month. During the second quarter earnings call, Samsung stated that it expects HBM4 revenue to increase more than threefold quarter-over-quarter in the third quarter, with HBM4 revenue in the second half expected to exceed 60% of the company's overall HBM revenue.

According to a report by the Economic Daily, with the rising demand for AI and high-performance computing applications, TSMC is expanding its capacity for advanced processes below 3 nanometers at the request of its customers.Industry insiders are optimistic that TSMC's contribution from 3 nanometers in the second half of the year is expected to surpass that of 5 nanometers for the first time, estimating that the single-quarter output value in the third quarter will exceed 400 billion New Taiwan dollars, setting a new record, accounting for more than 30% of revenue, and positively impacting gross margin performance.

On September 3 local time, Tesla held a Cybercab launch event in Austin, Texas, USA. This vehicle is a mass-produced model designed for autonomous driving scenarios, eliminating the steering wheel, pedals, and traditional rearview mirrors, and adopting a two-seat layout, with driving entirely controlled by Tesla's autonomous driving software. The Cybercab was unveiled as a concept car at the "We, Robot" event in October 2024, with the first mass-produced vehicle rolling off the line at the Texas Gigafactory in February this year, and mass production starting in April, taking about 18 months from concept to mass production.The vehicle relies on 8 high-definition cameras to perceive the environment and makes driving decisions through an end-to-end neural network, without relying on lidar and high-precision maps. According to Tesla's data, as of September 2026, the global fleet's cumulative assisted driving mileage has exceeded 22.5 billion kilometers. In terms of energy efficiency, it is expected to travel 9.8 kilometers per kilowatt-hour, with an energy consumption of about 10.2 kilowatt-hours per 100 kilometers. The curb weight is 1412 kilograms, equipped with a 47.6 kilowatt-hour battery pack and a 163-kilowatt motor, with a laboratory range of 673 kilometers. After scaling up ride-hailing operations, the cost per mile can be reduced to $0.2, equivalent to about 0.84 yuan per kilometer.The Cybercab will be showcased in multiple cities such as Beijing and Shanghai starting in mid-September. This exhibition does not involve the sale of the vehicle in China, nor does it represent that Tesla will conduct autonomous ride-hailing commercial operations domestically.

Stablecoin payment infrastructure startup Diameter Pay has completed a $10 million Series A funding round, co-led by CMT Digital and Lightspeed Faction.Diameter Pay stated that since the beginning of this year, it has processed over $10 billion in payments by helping institutional clients access dollar accounts, payment channels, and stablecoin deposit and withdrawal pathways.

Tencent Hunyuan has released and open-sourced the next-generation large language model Hy4 preview. This model has a total of 770B parameters and 49B active parameters, with a context length exceeding 1M. It demonstrates strong capabilities in real productivity tasks such as coding, office work, and science, firmly placing it in the top tier of open-source models. Hy4 preview significantly expands in model size, context length, and data scale, and enhances real-world performance through high-quality data co-built with Tencent experts in software engineering, gaming, finance, security, and deep collaboration with products like WorkBuddy.In software engineering, the model enhances understanding, planning, debugging, and verification capabilities for long-range development tasks, enabling the construction of complex front-end projects like a Three.js miniature town from scratch. In game development, it supports generating playable prototypes from a single sentence and can complete a full demo in Unity. In smart office applications, it can handle complex financial audits, filtering, analyzing, and delivering from multiple documents. In scientific research, it achieves acceleration in tasks such as molecular dynamics simulations and has initially formed a recursive self-improvement feedback loop.Hy4 preview can be experienced in Tencent products such as WorkBuddy/CodeBuddy domestic and international versions, Yuanbao, ima, and can also be accessed via API calls through Tencent Cloud Tokenhub and OpenRouter. WorkBuddy/CodeBuddy will launch a limited-time free activity for two weeks. Since the reconstruction of the infrastructure, the Hunyuan large model has iterated a major version approximately every two months, continuously optimizing through a preview-first and formal version-following approach.

Yesterday, Ledger disclosed details of the LSB 023 security vulnerability on its official website. Some applications built on the Ledger Secure SDK may still receive new APDU commands during the user screen confirmation, leading to inconsistencies between the parameters displayed on the screen and the final signature parameters.In cases where the attacker controls the APDU communication between the device and the host, after the user confirms the operation shown on the screen, the device may generate a signature with different parameters. Ledger stated that it has fixed the issue through application-level checks and the SDK layer, and released Ledger Secure SDK v26.6.1 on August 21. The relevant applications have been rebuilt and released. Users need to update the applications through Ledger Live; merely updating the device firmware is insufficient to complete the fix. However, Ledger claims there is currently no evidence that the vulnerability has been exploited.

Google has confirmed that it is gradually promoting the "goto" redirect parameter in search result links. When users click on a link, they will first go through a redirect address with the google.com/goto parameter, rather than directly pointing to the target webpage. This mechanism increases the difficulty for automated tools to batch parse and scrape search results through server-side redirection, in response to scraping behaviors from third-party tools and AI companies.SEO monitoring platform Nozzle's data shows that this change has nearly 100% coverage in multiple residential IP service provider environments.

Apple's first 2-nanometer M6 chip officially debuts, equipped with a 12-core CPU, 12-core GPU, and dual 16-core neural network engines, with unified memory bandwidth reaching up to 170GB per second, initially featured in the new Mac mini. This chip is manufactured by TSMC using a 2-nanometer process and is the first to adopt gate-all-around (GAA) nanosheet transistors, making it the world's first consumer-grade 2-nanometer chip.The supply chain is focused on the subsequent high-end M series packaging architecture. Based on the Fusion Architecture of the previous generation M5 Pro, M5 Max, and the four-die design of M5 Ultra, Apple chips have transitioned from a single large die to a modular design. Future M6 Pro, Max, or Ultra are expected to enhance advanced packaging requirements such as SoIC-MH and WMCM, increasing interconnect density with SoIC-MH and integrating logic, LPDDR memory, and high-speed I/O with WMCM.TSMC is actively preparing for expansion, with Zhunan AP6 as the main mass production base for SoIC, Longtan AP3 upgrading to WMCM, and Chiayi AP7 taking on related capacity. Analysts estimate that WMCM's monthly production capacity will reach about 60,000 units by the end of 2026 and over 120,000 units in 2027. Equipment manufacturers such as Hongshuo, Junhua, Yinneng, and material manufacturers Changxing, Xinying Materials, and Yongguang are expected to benefit.

According to Jinshi reports, the dark side of the moon Kimi announced that the company's first-generation trillion-parameter multimodal model Kimi K2.5 will end its service at the end of this month.

According to LatePost, ByteDance is discussing a large model with training parameters exceeding 50 trillion, surpassing Alibaba's Qwen 3.8-Max (24 trillion) and Moonlight K3 (28 trillion), making it the largest known plan in the country so far. This plan is still in its early stages and does not guarantee a final release. The new model is intended to be led by Xiang Liang, head of Seed Foundation, in collaboration with Shen Ke, who is responsible for the pre-training data of large language models. Seed is reorganizing, dividing responsibilities, and allocating resources based on this.Two weeks ago, ByteDance founder Zhang Yiming held a company-wide meeting with Seed head Wu Yonghui. Zhang reassured the team that training large models is inherently difficult and that it is acceptable to lag behind for a period of time, hoping to aim for the upper limits of intelligence and join the world's top tier. He acknowledged that programming is a key direction at present, advocating for the integration of Volcano Engine, Feishu, and Doubao resources to build computational power and data advantages, while reminding not to be led by a single hot topic. He praised Seedance's differentiated leadership and clearly opposed distillation, believing it is difficult to truly surpass and that AGI barriers should be built from a more fundamental level, stating that the company will continue to increase investment in AI.In the past six months, Seed's multimodal performance has been outstanding, with Seedance 2.0, Seedream, and others driving Volcano Engine MaaS, but the market response to the language model Seed 2.0 has been limited, and its lagging coding capabilities have affected the revenue structure. ByteDance has hired Guo Daye at a high salary to specialize in coding and has consolidated related resources. In the face of the industry's general trend of increasing model sizes, ByteDance hopes to achieve a leapfrog advantage with a larger scale while promoting the elimination of horse racing and breaking down departmental walls to concentrate efforts on tackling challenges.

Park Chang-min from Samsung Electronics Semiconductor Research Institute's Foundry Process Development Team stated at the "2026 Next Generation Lithography + Patterning Academic Conference" that the company plans to apply High-NA EUV technology starting from the 1-nanometer (A10) process, aiming for mass production around 2030. Samsung is currently still using the existing 0.33 NA EUV multiple patterning scheme for processes at 2 nanometers and below, with plans to achieve mass production of 1.4 nanometers by 2029, and then advance to 1 nanometer.High-NA EUV will increase the lens numerical aperture (NA) from 0.33 to 0.55, providing higher resolution, which can simplify ultra-fine processes that originally required multiple patterning into single patterning, helping to reduce manufacturing costs, improve production efficiency, and enhance design flexibility. However, this technology is highly challenging, equipment is expensive, and it requires simultaneous upgrades in supporting material technologies. Park Chang-min expects that the 1.4 nanometer and 1 nanometer processes will still primarily use 0.33 NA EUV multiple patterning, after which High-NA patterning will become the core technology of the new generation processes.

According to the Financial Times, as Chinese companies continue to narrow the gap with the top AI labs in the United States, ByteDance is training an AI model that may be close in scale to Anthropic's advanced Mythos system. According to three insiders, this Chinese tech giant is currently in the early stages of training a large-scale model, which may have a parameter count reaching the trillions, approximately three times that of the largest model released in China, the Kimi K3 model under the Moonlight brand.One insider stated that ByteDance's model is currently undergoing pre-training, a phase that typically takes three to six months. If all goes well, it will then undergo fine-tuning and be released. The specific scale of the model will be determined later. Anthropic has not publicly disclosed the parameter scale of its models, but industry estimates suggest that its advanced Mythos 5 model has a parameter count of about 8 trillion, while Fable 5 has around 5 trillion.It is important to note that the number of parameters determines the model's basic capacity for storing information or memory limit, but the actual capabilities of the model also depend on other factors, including data quality and training methods.

According to a report by LatePost, ByteDance is discussing training a model with a parameter scale exceeding 50 trillion, surpassing Alibaba's Qwen3.8-Max (24 trillion parameters) and the K3 from Dark Moon (28 trillion parameters), making it currently the largest known model in terms of parameter scale in the country. Generally speaking, the larger the model scale, the higher the level of intelligence.However, this plan is still in the early stages and does not guarantee that the model will ultimately be released. The new model will be led by Xiang Liang, head of SeedFoundation, in collaboration with Shen Ke, who is responsible for the pre-training data of large language models. To this end, Seed is reorganizing, defining responsibilities, and allocating resources.

The Web3 fitness protocol BRT.Fit has released a transparency statement, currently in the product testing and ecosystem validation phase. Application functions and validation methods may be adjusted based on testing results. The test token tBRT is only used for economic model validation and does not represent the official token value, nor does it constitute an investment product or guarantee of future returns. Its "mandatory exercise protocol" is an incentive and commitment mechanism that users voluntarily join, helping to develop exercise habits through verifiable data, without physical compulsion.After the official mainnet launch, core functions such as token staking, rewards, and burning will be executed through on-chain smart contracts, with contract addresses and audit reports to be disclosed gradually. The project plans to support mainstream wearable devices such as Apple Watch, WHOOP, and Xiaomi, Huawei, and will continue to optimize the anti-cheating system. The project team reminds that digital assets carry significant volatility risks, participation does not guarantee returns, and users should refer to official announcements, remaining vigilant against counterfeiting and phishing activities.

According to Xinhua News Agency, as the 2026 World Artificial Intelligence Conference and the High-Level Meeting on Global Governance of Artificial Intelligence are about to be held, the new generation model Kimi K3 was released on the 16th, with a parameter scale of 28 trillion. This is currently the largest open-source model in the world by parameters, marking a new step in the development of artificial intelligence models in our country.Beijing Moon's Dark Side Technology Co., Ltd., which released the model, introduced that Kimi K3 natively supports visual understanding, has a context window of 1 million tokens, and is optimized for complex task scenarios such as software engineering, knowledge work, deep research, and multimodal understanding, further enhancing the processing capability of large models for complex tasks. It is understood that in evaluations, Kimi K3's comprehensive intelligence level is close to that of the world's leading closed-source models. A relevant person in charge of Moon's Dark Side stated that in the process of training the new generation foundational model Kimi K3, they used a self-developed underlying model architecture and accumulated a complete set of scientific model training methods.

According to the Financial Times, informed sources reveal that the Chinese AI unicorn company Moonshot AI plans to release a new large language model, Kimi K3, in the near future. This model has between 20 trillion to 30 trillion parameters, making it the largest AI model in China by parameter scale, and its performance is expected to surpass the flagship model Claude Opus 4.8 from Anthropic in mainstream benchmark tests (industry speculation suggests its parameter count is around 15 trillion to 20 trillion).Unlike the currently mainstream closed-source and expensive cutting-edge large models in the United States, Kimi K3 will be available as an open-weight model for users to download and modify for free, which may create competitive pressure for leading American labs like OpenAI and Anthropic. Currently, due to the rising service fees for large models in the U.S. (for example, Anthropic has announced a 50% price increase for Opus 4.8 in September), some overseas companies have begun to shift towards using more cost-effective Chinese open-source models.In terms of the capital market, informed sources indicate that Moonshot AI is preparing for a new round of financing, with the latest valuation expected to reach approximately $31.5 billion. Meanwhile, the valuations of other AI giants in China and the U.S. are also rising; DeepSeek is starting a new round of financing with an estimated valuation of about $71 billion, while Anthropic and OpenAI have reached valuations of $965 billion and $852 billion, respectively, in their latest round of financing. In response to the aforementioned release and financing rumors, Moonshot AI has currently declined to comment.

Global chip foundry giant TSMC announced its financial report for the second quarter of 2026. Benefiting from the strong demand for advanced process chips driven by global AI infrastructure development, TSMC's performance this quarter significantly exceeded market expectations. During the period, it achieved revenue of NT$1.27 trillion (approximately US$40.2 billion), a year-on-year increase of 36%; net profit reached NT$706.6 billion (approximately US$22 billion), a year-on-year surge of 77.4%, far exceeding the market's previous estimate of NT$623.7 billion. In addition, the company's gross margin for the quarter reached 67.7%, and the operating margin was 60.3%, both better than expected.In terms of process structure, advanced processes (7 nanometers and below) contributed a total of 77% to the total wafer revenue this quarter. Among them, the 3-nanometer and 5-nanometer processes accounted for 30% and 33%, respectively, while the 7-nanometer process accounted for 11%. Notably, TSMC's newly shipped 2-nanometer advanced process recorded revenue for the first time, accounting for 3%.Looking ahead, TSMC confirmed that its capital expenditure for 2026 will approach a record US$56 billion and plans to invest approximately US$26.5 billion in its advanced manufacturing park in Arizona, USA. TSMC CEO C.C. Wei stated that the current pace of capacity expansion still lags behind demand, and the situation of supply not meeting demand is expected to continue for several years. Meanwhile, despite TSMC's strong performance, the market remains somewhat cautious and concerned about whether the massive AI investments by tech giants can translate into actual returns and the medium- to long-term competitive landscape.