BlockBeats News, July 28th: Since last night, a series of "major negative news" related to global semiconductor transactions has emerged, causing the U.S. semiconductor sector to come under pressure. Nvidia fell by about 5%, the Philadelphia Semiconductor Index dropped by 2.23%, and the Nasdaq experienced a slight decline.

During the Asian trading session, AI hardware assets in South Korea, Japan, and Hong Kong continued to decline. The South Korean KOSPI index fell by over 10% intraday and triggered circuit breakers twice, with heavyweight stocks SK Hynix and Samsung Electronics experiencing significant declines. In the Hong Kong market, leveraged products related to the semiconductor sector saw further intensified declines, with both the Southbound 2x long Hynix ETF and the 2x long Samsung Electronics ETF falling by over 20%.

The "major negative news" that triggered this round of selling has four main clues: first, concerns about the progress of China's domestic semiconductor production causing worries about competition; second, Nvidia's AI infrastructure-related transactions reigniting doubts about "circular financing"; third, the semiconductor sector had previously seen excessive gains, leading to profit-taking as funds concentrated at high levels; and fourth, the open-sourcing of Kimi K3 once again raised questions in the market about the high spending of U.S. and South Korean tech giants.

On the news front, according to The Information, a Chinese state-backed enterprise has begun mass-producing domestically developed immersion DUV lithography equipment, with plans to manufacture around 5 units in 2026 and expand to around 20 units in 2027. While this scale is still significantly smaller than ASML's delivery of 131 immersion DUV systems last year, the entry of domestic DUV production is seen by the market as a key breakthrough in China's chip supply chain localization. As a result, ASML plummeted by 5.80% on Monday, with U.S. storage and equipment chains weakening simultaneously, as SanDisk fell by 11.02% and Western Digital fell by 4.21%.

However, several institutions believe that the market's reaction may be too extreme. JPMorgan Chase pointed out that Chinese domestically produced immersion DUV equipment is still in the early stages of small-scale production, and its performance, reliability, and large-scale production capabilities are yet to be validated, indicating a long way to go before truly replacing ASML's equipment. Samsung Securities also believes that Chinese AI chips and server DRAMs will have difficulty entering the U.S. data center ecosystem in the short term, limiting the actual impact on the current AI semiconductor cycle. Citrini analyst Jukan stated that the substantial incremental value disclosed in the related reports is limited, and the sell-off reaction of ASML and other semiconductor equipment stocks is excessive.

Another source of pressure comes from the Chinese storage industry. ChangXin Memory Technologies soared by 466% on its first day of listing, becoming the largest A-share listed company in China by market capitalization, a development interpreted by the market as possibly accelerating China's self-sufficiency in the storage industry. For global storage industry leaders such as SK Hynix, Samsung Electronics, SanDisk, and Western Digital, the presence of well-capitalized Chinese competitors implies a potential shift in the long-term supply landscape, prompting a reevaluation of the previously bearish narratives in today's market.

Meanwhile, concerns over credit risk stemming from NVIDIA's $750 billion AI infrastructure partnership are spilling over from the bond market to the stock market. Bloomberg reported that NVIDIA is said to be discussing providing up to $250 billion in guarantees to assist OpenAI in leasing computing power for its U.S. data center project; the company also disclosed a collaboration with the SK Group, the parent company of SK Hynix, exceeding $500 billion. ICE Data Services data shows that NVIDIA's 5-year CDS spiked by around 14 basis points at one point, reaching as high as approximately 82 basis points per year, marking the largest intraday increase since the commencement of trading in the relevant contract.

CDS can be simply understood as "debt default insurance." When CDS prices rise, it usually indicates that the market believes the related company's debt risk or potential payment pressure is increasing. For NVIDIA, the credit market is reassessing the balance sheet pressure that could arise from its large-scale guarantees, customer financing, and AI infrastructure partnerships. Hideyuki Ishiguro, Chief Strategist at Nomura Asset Management, stated that the rising credit risk for NVIDIA is being viewed by investors as a negative signal.

Pressure is also brewing at the AI model level. Kimi K3 from the Dark Side of the Moon unveiled its model weights on July 27, which is widely seen in the market as another significant advancement in China's open-source AI models. Publicly available information indicates that Kimi K3 is a 2.8 trillion parameter-level model that supports long-context, multimodal, and agent capabilities, emphasizing the ability to approach cutting-edge models at a lower cost of use. Following DeepSeek, China's high-performance open-source models have once again entered the global investor spotlight, prompting the market to reevaluate the valuation basis of U.S. closed-source models, cloud providers, and the AI hardware chain.

For the U.S. stock market, the pressure from Kimi does not directly stem from the model itself but from the reinforcement of the narrative of "low-cost, high-performance AI." If model advancement no longer relies solely on larger-scale GPU clusters and higher capital expenditure, investors will naturally question whether the marginal returns on expanding data centers by companies such as Microsoft, Meta, Amazon, and OpenAI will decrease, and whether the high valuations enjoyed by NVIDIA, Broadcom, AMD, and the storage and hardware chain can be sustained. In other words, the open-sourcing of Kimi has deepened the market's concerns about AI capital efficiency and has put the trading logic of "the more computing power, the better" up for repricing.

More importantly, there has been a shift in the market's focus on AI trades. Previously, the surge in tech stocks was mainly driven by "high growth in computing power demand"; now, with the emergence of low-cost open-source models like Kimi, NVIDIA's customer financing arrangements, and massive data center capital expenditures all occurring simultaneously, investors are beginning to question whether AI investments can truly translate into sufficient cash flow, and how much of customer procurement demand relies on supplier financing support.

Macro-level pressure is also building. This week, both the Federal Reserve and the Bank of Japan will announce their interest rate decisions, stoking concerns about a "double Fed-Japan hawkish" stance. The Federal Reserve will hold its monetary policy meeting on July 28-29, with the current federal funds target range at 3.50%-3.75%. While most traders still expect a hold steady approach, the market has priced in about a one-third probability of a surprise rate hike. Goldman Sachs stated that the outcome of this Fed meeting is "exceptionally uncertain." Against the backdrop of AI capital expenditure, oil prices, and tariffs which could still boost inflation, if the Fed sends a stronger signal, the valuation pressure on high-growth tech and semiconductor stocks will further escalate.

On the Bank of Japan side, the meeting on July 30-31 is expected to keep the policy rate unchanged at 1.00%, but the market is more focused on whether it will signal future rate hikes. Reuters reported that the Bank of Japan may hold steady this week while leaving room for further rate hikes to address pressure from a weak yen, energy prices, and inflation expectations. The Bank of Japan previously raised rates by 25 basis points to 1.00% in June, the highest level since 1995. If the Bank of Japan continues its hawkish stance, the stability of the yen carry trade will be tested, and global risk assets may face additional deleveraging pressure.