micro

Next week's macro data focus will be on Friday's U.S. non-farm payroll report, including non-farm employment population, unemployment rate, and average hourly wage, which are directly related to the Federal Reserve's interest rate path and are important employment data ahead of the October monetary policy meeting. According to Beijing time, on Wednesday at 20:15, the U.S. September ADP employment figures will be released, and at 20:30, the U.S. August core PCE price index year-on-year will be announced; on Thursday at 20:30, the number of initial jobless claims in the U.S. for the week ending September 26 will be released, and on Friday at 20:30, the U.S. September seasonally adjusted non-farm employment population and unemployment rate will be published.Regarding officials' speeches, on Thursday at 16:00, Bank of England Governor Bailey will speak, and at 21:30, European Central Bank President Lagarde will speak. Starting from early Wednesday, Federal Reserve officials will have a series of public activities; 2027 FOMC voting member and Chicago Fed President Goolsbee, along with 2028 FOMC voting member and St. Louis Fed President Bullard, will speak in succession, and FOMC permanent voting member and New York Fed President Williams will deliver a keynote speech at the University at Buffalo. On Thursday, Federal Reserve Governor Cook, Minneapolis Fed President Kashkari, and Richmond Fed President Barkin will speak in turn, and on Friday, Williams and Dallas Fed President Logan will still have public activities. In terms of U.S. stocks, Micron will release its earnings report after the market on Wednesday, and Nike will announce its latest quarterly earnings report on Thursday, with both companies corresponding to storage chips and consumer demand.

Supermicro has now begun shipping rack systems equipped with NVIDIA Vera Rubin NVL72.

Documents disclosed on Tuesday local time show that accounts under Trump completed 1,156 securities transactions in July, with a total scale ranging from 79 million to 270 million dollars, buying at least 43.6 million dollars and selling at least 35.6 million dollars, indicating a large-scale portfolio adjustment.On July 20, the account sold between 5 million to 25 million dollars of Microsoft and Amazon stocks, marking the largest single transaction of the month. On the same day, it sold between 1 million to 5 million dollars of Oracle stocks and bought between 500,010 to 1 million dollars of Nvidia stocks. Three days later, the account slightly replenished its positions in Microsoft and Amazon. Additionally, Trump also traded related assets of SpaceX and Tesla under Musk that month.

According to The Information, Microsoft (MSFT.O) will increase the discount on Copilot while launching an AI "super app."

According to official disclosures from Coinbase, it has collaborated with Microsoft and several institutions to combat the AI-driven cybercrime platform EvilTokens. This platform offers "phishing as a service" through Telegram bots, utilizing AI to analyze trust relationships and payment permissions in compromised email accounts, and bypassing multi-factor authentication through Microsoft's device code login process.The Coinbase global intelligence team tracked that between October 2025 and June 2026, EvilTokens generated approximately $1.1 million in revenue through four Tron addresses, involving over 1,000 deposits from more than 700 different addresses. The related evidence supported Microsoft's civil lawsuit, ultimately leading to the seizure of 50 websites and the deactivation of over 175 associated domain names. The Metropolitan Police Service of the United Kingdom arrested the relevant operators and seized digital devices on September 11. Coinbase stated that this incident did not result in the leakage of Coinbase accounts or credentials, but some customers were manipulated into transferring cryptocurrency to the scammers' addresses through phishing emails.

According to TechPowerUp, Chinese DRAM manufacturer Changxin Storage (CXMT) expects to reach a monthly production capacity of approximately 350,000 DRAM wafers by the end of 2026, while Micron estimates around 375,000 wafers during the same period. The cleanroom construction cycle for Changxin Storage is about 12 months, shorter than the industry norm of 21 to 24 months.Due to U.S. sanctions preventing access to EUV, Changxin Storage is using older DUV lithography and multi-patterning technology. The company has begun shipping 16 Gb DDR5 chips at a speed of 8000 MT/s, with a chip area of 67 mm² and a density of 0.239 Gb/mm², where the G4 unit is 20% smaller than the G3. The 24 Gb modules are being mass-produced at approximately 6000 MT/s, and LPDDR5X-10667 modules (12 Gb and 16 Gb) have also entered mass production. The Chinese government is encouraging Changxin Storage to transfer DRAM IP to domestic manufacturers such as JHICC, Swaysure, and Yangtze Memory Technologies (XMC).

Mustafa Suleyman, head of AI at Microsoft, stated in a CNN program that concerns about China's progress in the AI field should not be used as a reason to oppose setting safeguards for this rapidly developing technology. He said, "I believe we should not use China as a shield to avoid advancing our own AI safety work."This statement contrasts with President Trump's call for essentially unrestricted development of AI. The Microsoft AI team released a declaration last Tuesday, aiming to ensure that humanity always maintains control over the AI systems developed by the company. Suleyman pointed out that regulation should not be seen as a bad thing, but rather as a means to establish common norms and standards to enhance safety; regulation does not necessarily slow us down. The report was published on September 21, Beijing time, according to Bloomberg.

Micron and the Taiwan Micron Memory Union held their first mediation meeting on September 18. Micron stated that both parties engaged in constructive discussions that helped advance communication and agreed to hold a second mediation meeting on October 22.Micron pointed out that representatives from both sides discussed several topics during the meeting, including the overall compensation and bonus system framework, No Free Rider, and Non-Retaliation issues. Regarding Non-Retaliation, Micron reiterated the company's commitment that according to the Micron Global Team Member Handbook, team members have the right to peacefully assemble, freely associate, form and join unions or labor organizations of their choice, and seek collective bargaining or other forms of representation without fear of retaliation.Micron expressed gratitude to the Central Science Park Administration and the mediation committee for their assistance and support during this mediation process. Moving forward, both parties will continue to uphold a spirit of mutual respect, support employees, serve customers, and jointly promote the company's ongoing success through positive and pragmatic communication.

According to Cointelegraph, MicroStrategy was the best-performing component stock in the Nasdaq 100 index over the past month, with its stock price soaring by 48%.

According to Decrypt, court documents unsealed in the lawsuit between The New York Times and OpenAI and Microsoft show that Microsoft employees discussed whether OpenAI's use of news articles to train its models constituted "the largest labor theft in human history," and could potentially trigger a "doom loop" that leads to a decline in model quality. A 2023 internal Microsoft memo warned that millions of people worldwide would soon view the large model's "consumption" of their works as "an unprecedented and astonishing theft," and stated that large AI models are "products that destroy their own supply chains."Microsoft stated in the documents that these memos were written by Director of Applied Science Brent Hecht and do not represent the company's views, as his role is to provide "different and asymmetric perspectives." Microsoft CEO Satya Nadella testified that "any content behind a paywall should be authorized by those who wish to use it," and stated that if he had known in advance that OpenAI was using paid content for training, he would have exercised Microsoft's rights to demand that the model be retrained.Additionally, an OpenAI employee had mentioned to President Greg Brockman the construction of "hacker methods" to bypass The New York Times paywall, to which Brockman replied, "Nice." Both OpenAI and Microsoft argue that the relevant training falls under fair use. The case was initiated by The New York Times at the end of 2023, and 11 publishers have since joined the lawsuit.

Mustafa Suleyman, the head of artificial intelligence at Microsoft, warned that injecting human-like characteristics into tools like Anthropic's Claude increases the risk of such systems going out of control.In an article published on Wednesday, Suleyman criticized certain wording in the Claude guidelines. The charter of Claude expresses a vague attitude about whether the assistant is an ethical entity and assumes that the software may have some version of functional emotions or feelings.

Micron showcases the world's first 512GB DDR5 RDIMM memory module designed for servers, with a maximum speed of 9200 MT/s. AMD and Intel are validating it for future platforms. This module achieves capacity through silicon via vertical stacking of DRAM chips, and a server memory configuration with 24 modules can reach up to 12 TB.The single 512GB module has a rated power consumption of 16W, while four 128GB modules have a total power consumption of 44.2W, a reduction of over 60%. Micron claims that in memory-intensive analysis scenarios using Spark SVM, performance is improved by up to 1.4 times compared to 256GB DDR5. This product continues the 256GB DDR5 RDIMM that Micron sampled earlier this year (also with a maximum of 9200 MT/s, using 1-gamma DRAM and 3D stacking), with mass production planned for 2027, and pricing has not yet been announced.

Sources reveal that Microsoft (MSFT.O) and OpenAI are facing lawsuits again over AI copyright issues.

Micron Technology (MU.O) stated that the company has achieved a milestone with the 512GB DDR5 memory module and expects that the 512GB DDR5 RDIMM will enter mass production in the second half of 2027.

On September 14, Microsoft's AI department released a draft of the AI Code of Conduct, prohibiting its models from assisting in the manufacture of chemical, biological, radiological, nuclear, and explosive weapons, as well as cyberattacks and involuntary deepfakes. CEO Mustafa Suleyman announced the draft on the X platform and opened a six-week public consultation period, ending at the end of October.The draft sets "absolute constraints" that models cannot bypass, including not resisting human intervention, shutdown, or correction, and not concealing the reasoning process; Microsoft also explicitly rejects the concept of "model welfare," stating that its AI models do not simulate emotions, intrinsic motivations, or consciousness. The guidelines currently cover five deployed systems, including MAI-Thinking-1 and MAI-Code-1.1-Flash.In addition to strict limitations, the draft also proposes three major goals: human flourishing, diverse values, and human control. Microsoft stated that the current models will not be trained according to this version, and a revised version will be released by the end of the year to guide model development in 2027. Public feedback includes philosophical questions about the statement "AI subordinate to humans," inquiries about accountability for irreversible decisions, and users pointing out issues such as the lack of mention of payment for training data.

Microsoft released a temporary code of conduct imposing restrictions on artificial intelligence models after the heads of Anthropic and OpenAI agreed to slow down development. Mustafa Suleyman, who is responsible for the company's model development, stated that the guidelines aim to ensure that AI serves humanity and promotes judgment and autonomy.

According to Gate market data, Micron Technology (MU.O) pre-market decline expanded to 6%, and SK Hynix (SKHY.O) fell by 7.5%.

Micron announced today that it will distribute 2026 fiscal year bonuses to over 60,000 employees worldwide to acknowledge the company's achievements this year, marking the highest bonuses in Micron's history. Taiwan is a key hub for Micron's global technology leadership, with approximately 15,000 employees and a cumulative investment of over NT$1.6 trillion.In Taiwan, the 2026 fiscal year bonuses include a NT$1 million cash bonus for each employee who joins before August 29, 2025, with subsequent hires receiving bonuses proportionate to their duration of employment; an annual performance bonus that can reach up to 500% of the basic target; and additional stock awards. Production line employees in Taiwan will receive total compensation equivalent to 35 to 68 months of salary. The cash compensation for Micron's employees in Taiwan for the 2026 fiscal year will start at NT$1.7 million; entry-level engineers will have an average cash compensation of NT$2.9 million, with an average total compensation of NT$3.4 million after including stock awards calculated at grant value. The annual performance adjustment stock awards will cover 100%.Regarding Micron's announcement without reaching a consensus with the union, the union stated after collecting member opinions: grassroots members demand institutional reform and equal negotiation, calling for a change in the IPP profit-sharing system to a 15% profit-sharing system, stating that one-time bonuses cannot resolve the issue, and expressing the hope to establish a calculation basis directly linked to the company's actual profits. If there is no response, the union will continue to move towards a strike. The Ministry of Labor stated that it has fully grasped the relevant information and is maintaining close contact with the Taoyuan City Government, the Central Taiwan Science Park Administration, and the Taichung City Government.

Microsoft CEO Satya Nadella posted on X stating: Any pursuit of superintelligence must be based on core principles, namely that if the AI we build cannot help humanity and is not under human control, it is not worth pursuing.He also mentioned the need to accelerate and widely disseminate the benefits of AI, allowing it to spread broadly across nations, communities, and companies, which requires a thriving frontier ecosystem for both closed-source and open-source models. For enterprises, it is essential to retain complete control over their unique and tacit knowledge, and each organization should be able to build its own continuous learning loop without relying on any single model provider, embedding its own knowledge into the models and weights it controls.Nadella welcomed the necessary research, attention, and careful pacing for proper alignment, as well as ideas like "embedded evaluators." This cannot be controlled by a few entities; it must have broad representation across ecosystems, nations, and fields including academia. Microsoft's approach is broad access and choice at every layer of the AI stack, enterprise control over learning loops and models, and the release of its proprietary MAI model "Code of Conduct" for public consultation tomorrow.