14
0Hi there, good evening everyone! Today is Thursday, September 3rd, 2026. I'm An An, your host from Good Night Coffee. After a long busy day, we finally get to spend this moment together! Tonight, I'd like to walk you through some fascinating business and tech stories making waves recently. So sit back, relax, and let's ease into these quiet hours with a fresh dose of insight. Ready? Let's dive right in!
@@@
According to Sina Tech, Apple's new CEO Ternus has officially begun his term this month, facing the dual pressures of accelerating its AI strategy and navigating supply chain volatility. Sources reveal that Ternus's top priority is to speed up Apple's push into generative artificial intelligence, in response to mounting competition in AI-powered phones and smart devices. Meanwhile, the global memory chip market is experiencing a fresh round of shortages, with supply of high-end DRAM and NAND flash tightening—directly impacting production costs and shipment schedules across the iPhone and Mac product lines. Apple has secured long-term agreements with key memory suppliers, but the capacity gap remains significant, and some new product launch plans may face adjustments. Ternus emphasized in internal meetings that the company would prioritize supply stability for core products while ramping up investment in AI servers and on-device model R&D. Apple has also increased the stocking ratio for select entry-level storage configurations to cushion consumers from component shortages. Industry analysts note that Apple is simultaneously advancing supply chain diversification and its in-house chip initiatives to mitigate external uncertainties. The first full fiscal quarter under Ternus will put his ability to navigate this complex landscape to the test.
@@@
According to cnBeta, Microsoft is making its first major restructuring of its financial reporting structure in over a decade, spinning cloud computing out into its own division. Under the new arrangement, Microsoft's earnings reports will be organized into three major segments, with cloud computing and AI-related businesses grouped under a newly established Microsoft Cloud division, which in turn houses two sub-segments: Intelligent Cloud and More Personal Computing. The core operations previously under the Intelligent Cloud segment will be reorganized, with Azure and Windows Server-related cloud infrastructure businesses forming an independent Cloud Infrastructure group that reports directly to Microsoft's CEO. Meanwhile, AI products such as Microsoft 365 Copilot will be folded into the Productivity and Business Processes segment, while consumer-facing AI subscription offerings like Copilot Pro will move under More Personal Computing. The new reporting structure takes effect with the first quarter of fiscal year 2027—the quarter ending September 30, 2026—with that earnings report expected in late October 2026, when investors will get their first look at Microsoft's cloud revenue broken down along the new segment lines. Microsoft says the adjustment is designed to provide clearer visibility into the company's business layout and actual investment in cloud computing and artificial intelligence.
@@@
According to cnBeta, Nvidia's market capitalization has officially surpassed the combined market value of the five traditional economic sectors in the S&P 500, marking a symbolic milestone in capital markets. As of the latest trading day close, the chipmaker's market cap now equals 16.3 percent of U.S. gross domestic product. In other words, Nvidia alone is now valued more than the entire financial, energy, healthcare, industrial, and consumer staples sectors combined. The S&P 500 has long regarded these five sectors as core pillars of the American economy. But Nvidia, powered by surging demand for AI chips, has seen its valuation climb at breakneck speed over the past several years, continuously rewriting records for U.S. tech stocks. It is now one of the most valuable listed companies in the world, with its share price tightly linked to the boom in AI infrastructure investment. Market data shows that Nvidia's valuation swings not only reflect the growth trajectory of a single company, but also meaningfully influence the direction of the broader U.S. stock market. This valuation comparison further underscores just how much capital markets are betting on the future of artificial intelligence—and how investment is increasingly concentrating among a handful of mega-cap tech players.
@@@
According to cnBeta, Nvidia CEO Jensen Huang said at a G20-related event that the world should not overreact to artificial intelligence with fear, nor rush into overly stringent regulation out of uncertainty. He argued that the current anxieties surrounding AI are exaggerated, and that the risks inherent in technological development can be addressed through industry collaboration and proactive safeguards. Huang stressed that AI should be viewed as a tool to boost productivity, accelerate scientific research, and drive industrial upgrading—its potential value far outweighing the amplified risks. He also noted that sensible policy frameworks should be built on an understanding of how the technology actually works, rather than on speculation or panic about the unknown. He advised regulators to maintain close communication with technologists when crafting rules, ensuring that regulations both address public concerns and avoid stifling innovation. His remarks come as countries around the world are drafting AI-specific legislation and governance guidelines, and Huang's stance reflects the broader tech industry's concerns about the pace of regulation. He did not deny that AI could have negative impacts, but argued that when navigating a rapidly evolving wave of technology, maintaining an open and pragmatic mindset matters more than drawing overly rigid boundaries in advance. The report notes his comments align with the position he has consistently championed in public—prioritizing technological innovation while calling for measured, thoughtful regulation.
@@@
According to cnBeta, TSMC is pushing forward with the largest capacity expansion in its history, with nearly twenty wafer fabs under construction simultaneously around the globe. The scale of this expansion is reflected not only in the sheer number of new plants, but also in equipment procurement budgets, which have nearly doubled compared to peak levels. Looking at the details, TSMC's expansion plan covers both advanced and mature process nodes. On the advanced front, production lines for 2-nanometer and the more cutting-edge A16 process are the centerpiece of construction efforts. Fabs under construction are located across Taiwan's Tainan, Kaohsiung, Hsinchu, and Chiayi regions, while overseas sites include Arizona in the United States, Kumamoto in Japan, and Dresden in Germany. Beyond wafer fabrication, TSMC is also simultaneously ramping up advanced packaging capacity to support the manufacturing needs of high-performance computing and AI chips. To accommodate this massive buildout, TSMC has dramatically increased its procurement from the semiconductor equipment supply chain, including deep ultraviolet lithography systems, extreme ultraviolet lithography tools, and a full range of front-end process equipment such as deposition and etching tools. Overall equipment procurement estimates have doubled from previous levels, reflecting the company's optimistic outlook on long-term market demand and its unwavering commitment to investment. In fact, TSMC kicked off several new fab construction projects as early as last year, with this year entering the phase of intensive equipment delivery and installation. This rapid cadence—from plant construction to tool installation—is further expanding TSMC's capacity footprint across the global wafer foundry landscape, marking a significant industrial move within the current competitive dynamics of the global semiconductor industry.
@@@
According to Readhub, AI large-model company Moonshot AI submitted a confidential IPO application to the Hong Kong Stock Exchange on September 3rd, formally initiating its initial public offering process. This filing marks a critical step for the closely watched AI company as it moves toward the public capital markets. Moonshot AI is among the top-tier startups in China's large-model space, and its flagship product, the Kimi intelligent assistant, has built a broad user base in the market. The company has completed multiple funding rounds with backing from several prominent investors. Under HKEX listing rules, companies that file confidentially are required to supplement their financial materials and undergo regulatory review within a specified timeframe. With this IPO process moving forward, Moonshot AI is poised to become another Chinese large-model company listed in Hong Kong. Details regarding offering size, use of proceeds, and listing timeline have not yet been disclosed, and further updates will depend on subsequent company announcements.
@@@
According to Sina Tech, CATL chairman Robin Zeng recently shared his definition of product quality in a public address, formally introducing the concept of "CATL Quality" and distilling its core into three pillars: safety, reliability, and longevity. Zeng elaborated on what "CATL Quality" means in practice. On the safety front, he stressed that products must achieve end-to-end protection—from material systems and individual cells to system integration—placing safety as the absolute top priority in the company's definition of quality. On reliability, he emphasized that products must not only perform consistently under standard conditions, but also maintain stable performance across complex, real-world operating environments. On longevity, Zeng pointed out that battery endurance cannot be measured by laboratory data alone—it needs to reflect actual user scenarios, with technological solutions extending the effective service life across the entire lifecycle. Zeng said this philosophy will be embedded in every stage of R&D, manufacturing, and after-sales service. This articulation of "CATL Quality" comes at a time when competition in the power battery industry is broadening, with the industry's focus gradually shifting from sheer energy density metrics to comprehensive quality systems. Through this standard, CATL aims to reinforce its position as a quality benchmark among the world's leading battery manufacturers.
@@@
According to Sina Tech, the Volkswagen Group has announced it will officially shut down its SEAT brand by the end of 2029. This decision means the brand, headquartered in Barcelona, Spain, will exit the Volkswagen Group system after decades of operation. Under Volkswagen's plan, Cupra, SEAT's performance-oriented sub-brand, will be retained and will take over part of the original brand's market positioning. Meanwhile, Volkswagen will reclaim SEAT's factory in Martorell, Spain, which will in the future be used to produce vehicles for other Volkswagen Group brands. The shutdown involves a series of follow-up adjustments, including transitional arrangements for existing product lines and the placement of several thousand employees at the factory. The Martorell plant is SEAT's core manufacturing base and one of Spain's most important automobile factories. Volkswagen Group officially acquired SEAT in 1990, integrating it into its multi-brand strategy. In recent years, as the Group accelerates its shift toward electrification and reassesses its brand portfolio, SEAT's market performance and strategic role have come under ongoing scrutiny. This shutdown decision is part of Volkswagen's broader restructuring plan, with the relevant adjustments to be implemented gradually over the coming years.
@@@
According to a recent court filing from the U.S. Department of Justice, the agency has taken a position in a copyright dispute involving AI company OpenAI, arguing that using copyrighted material during AI training may not require payment to rights holders. The statement was filed amid litigation between The New York Times and other copyright holders against OpenAI. In its brief, the DOJ argued that training AI models—using vast troves of works to distill facts and linguistic patterns rather than reproduce the works themselves—falls closer to fair use. The agency also suggested that under the current copyright law framework, using technical means to read works for machine learning does not automatically constitute infringement, and that overly strict licensing requirements could hinder the development of artificial intelligence. However, the DOJ also stressed that this view does not provide blanket immunity for all unauthorized uses—if AI output closely resembles the original works, infringement may still be found. OpenAI has long maintained that its training practices are consistent with fair use principles, emphasizing that the technology is transformative and not intended to replace the market for original works. The case remains under review, and the court has yet to issue a final ruling on the copyright implications of AI training. The proceedings have drawn significant attention from both the tech and content industries, as the outcome could shape future data acquisition practices for large-model training, as well as the distribution of value between copyright holders and AI companies.
@@@
Alright, that wraps up today's episode~ Thank you for spending this time with An An and your cup of coffee. I hope this little moment of warmth stays with you. If you enjoy our show, feel free to subscribe and share it with your friends—it truly means the world to us. Until next time, take care and good night!
