Shield: The KiloEx Attacker Tagged Address has returned approximately 1.4 million USDT to KiloEx

Citigroup has cut Microsoft’s target price, citing the company’s AI investments are putting greater pressure on short-term free cash flow and profit margins, while maintaining a Buy rating. The bank lowered Microsoft’s target price from $620 to $570. It noted that Microsoft remains one of the clearest players in the commercialization of AI software and cloud computing, but the pace of capital expenditure on Azure data center construction, GPU procurement, and AI infrastructure expansion may continue to weigh on investors’ cash flow expectations in the coming quarters. The bank still believes Copilot, Azure AI services, and enterprise customer migrations will underpin long-term revenue growth. The main point of divergence lies in the timeline: the market was previously willing to price in AI earnings in advance, but is now starting to demand that Microsoft prove these expenditures can translate into profits faster. Mizuho and Wells Fargo have also recently cut Microsoft’s target prices, while maintaining positive ratings. The shared view among multiple institutions is that Microsoft’s AI assets remain attractive, but its stock price needs to digest near-term pressures from high capital expenditure, the AI-driven reshaping of its traditional software business, and a reassessment of Azure’s growth expectations.

According to South Korean media reports, Financial Services Commission Chairman Lee Won-yeon stated that regulators are developing a comprehensive plan for single-stock leveraged ETFs and plan to announce it soon. He added that regulators are closely assessing the impact of such products on recent volatility in South Korea’s stock market, but remain cautious about measures like trading halts, noting that potential market side effects need to be fully considered. The report noted that South Korean President Lee Jae-myung previously instructed financial regulators to quickly formulate supplementary regulatory measures for single-stock leveraged ETFs of companies including Samsung Electronics and SK Hynix. Launched in South Korea in May this year, these products offer 2x leverage on the daily price fluctuations of individual stocks and are classified as high-risk investment products.

According to Hyperinsight monitoring, yesterday midday, a whale with an address starting with 0x8af chased to build a long position in SK Hynix (SKHX) when the stock rebounded to around $1,440. The whale opened a total of 301 long orders in a short period, trading 6,008 contracts at an average price of approximately $1,440, with a total position size of about $8.6516 million. The entry price was roughly $1,439.91, precisely at the highest tier of this rebound. SKHX subsequently plunged sharply; as of press time, it is trading at around $1,237, down roughly 14% from the whale’s cost line. Focusing solely on the SKHX long position, the address currently holds a 10x full position, with its nominal size reduced to approximately $7.453 million, an unrealized loss of ~$1.199 million, and a return rate of -138.6%.

Bank of Korea Governor Hyun Song Shin announced that the central bank will raise the interest rate on its special loan program.

According to official announcements, Coinbase has announced it will add support for perpetual contracts on the Roundhill Memory ETF (DRAM) and Direxion Daily Semiconductor Bull 3X Shares ETF (SOXL).

Market research firm Citrini Research’s latest forecast indicates that by 2030, the global DRAM market will still face a massive supply gap, estimated at 28.7 EB, accounting for roughly 18% of total demand that year—compared to this year’s global total production capacity of around 40 EB. According to data from Citrini researcher Zephyr, global DRAM (including HBM) demand is projected to hit 157.5 EB in 2030, while supply capacity will only reach approximately 128.8 EB. General-purpose DRAM will be the biggest bottleneck: its annual supply is estimated at 91 EB, but demand stands at 120 EB, with the gap ratio expected to widen from the current 18% to around 25%. The report stresses that even as Samsung, SK Hynix, Micron, and Chinese manufacturers continue to expand production, new capacity may be quickly absorbed by surging AI demand. The core driver of the supply-demand imbalance is the boom in AI infrastructure: large model training and inference, plus HBM becoming a core component of AI accelerators, have jointly pushed up demand for traditional server DRAM. In a tight balance scenario, DRAM’s average selling price (ASP) is likely to stay high long-term, projected at $1.5–$2 per Gb, with memory costs for servers, PCs, and consumer electronics continuing to face pressure.