Societe Generale: The US dollar is expected to rise to 103.6 by the end of the year, while the euro will fall to 1.11 against the dollar.
Related
Video | Morgan Stanley's Chief U.S. Equity Analyst: Chip Stocks May Correct by Over 30%, But S&P Still Targets 8000 Points by Year-End
On July 6th, Morgan Stanley's chief U.S. equity analyst, Mike Wilson, stated in an interview with Bloomberg that while chip stocks remain core assets in the AI wave, some are severely overbought, and a short-term correction of 30% to 40% is entirely possible. However, he believes this does not signify the end of the U.S. stock market bull run, and the S&P 500 is expected to continue rising as funds rotate into other sectors. He maintains his prediction that the S&P 500 will break through 8000 points by the end of the year, believing that the core logic supporting the market still comes from corporate earnings growth.
Opinion: Open source models account for only 10% of enterprise large-scale model spending, but mature production environments will be dominated by open source models.
According to Beating's monitoring, while public opinion often touts that open-source large models are dominating everything, enterprise spending data presents the opposite picture. Jesse Zhang, co-founder and CEO of Decagon, an enterprise-level AI customer service platform, points out that the share of open-source models in total enterprise spending has now dropped to 11%. This decline stems from the fact that most enterprises' AI applications are still in the early, undefined exploratory stage, thus defaulting to reliance on closed-source models. However, he emphasizes that once application scenarios mature, open-source models will take over production environments with their advantages of extremely low latency and deep fine-tuning. In Decagon's own production environment, 90% of calls have already switched to open-source weighted models. The core driver of this transformation is interaction speed and customization capabilities, not cost savings. In customer service scenarios, a single conversation that takes 8 seconds to finish will completely destroy the product experience. Since leading closed-source labs do not allow fine-tuning of flagship models, and small closed-source models cannot be deeply customized, small-sized open-source models, through fine-tuning for specific tasks, have become the only option to support high-frequency real-time interactions. The future of enterprise AI will see a division of labor: leading closed-source labs will continue to dominate the exploration and discovery of new fields, while open-source weighted models will increasingly take over the actual production of mature businesses. Because model fine-tuning requires extremely high levels of data and talent, the migration from closed-source to open-source will be a slow process lasting several years, during which both will experience sustained growth.
Naver's $9.9 billion stock swap deal with Dunamu has been delayed again until the end of the year, while South Korea's digital asset law remains unresolved.
According to BlockBeats, on July 7, Naver Financial and Dunamu postponed the completion date of their full share swap transaction to December 31, marking the second delay in the deal. The deal to merge Dunamu, the operator of South Korea's largest crypto exchage Upbit, into Naver's financial sector was originally scheduled to close on September 30. Dunamu disclosed a new timeline on the 6th through corrections to its initial filing last November, but incomplete digital asset legislation and pending antitrust reviews remain major uncertainties. The company has postponed its extraordinary general meeting of shareholders from August 18 to November 19, and the shareholder confirmation date has been reset to October 22. Several government approvals are still required before the transaction can be completed, including approval from the Korea Fair Trade Commission (FTC) for the merger, approval for the change of Naver Financial's largest shareholder under credit information regulations, and acceptance of the filing for the change of Dunamu's largest shareholder under specific financial transaction information laws. Dunamu stated that progress at any of these stages could further extend the timeline or even cause the transaction to change. Dunamu also pointed out that the Digital Assets Basic Law, currently under consideration in Congress, is a real variable affecting the progress and outcome of transactions. As this bill is being enacted, regulators are also simultaneously considering implementing bank-style no-fault liability rules for exchanges, requiring platforms to compensate users for losses caused by hacker attacks.
Linkage Technology: First-half net profit expected to increase by 73%-139% year-on-year; semiconductor testing equipment industry continues to enjoy upward momentum; company has ample order backlog.
According to Mars Finance, LianDong Technology (301369.SZ) announced that it expects its net profit attributable to shareholders of the listed company for the first half of 2026 to be between RMB 21 million and RMB 29 million, representing a year-on-year increase of 73.39%-139.44%. The performance change is mainly attributed to strong demand from downstream application areas such as AI, high-performance computing, and new energy vehicles, leading to a continued upward trend in the semiconductor testing equipment industry. The company seized the strategic window of domestic substitution for semiconductor testing equipment, leveraging its technological accumulation and brand advantages to vigorously expand the market. During the reporting period, the company had a sufficient order backlog, achieving steady growth in operating performance. Note: The company's Q2 net profit is expected to be between RMB 18 million and RMB 26 million, compared to RMB 3 million in Q1. Based on this, Q2 net profit is expected to increase by 415%-649% quarter-on-quarter. (Cailian Press)
Bitget CEO sees Bitcoin near current levels at year-end, doubts US will buy BTC
Gracy Chen says macroeconomic uncertainty could keep Bitcoin within $10,000 to $20,000 of current levels, while US government purchases remain unlikely within the next two years.
Bernstein: The storage bull market could continue into 2027, but the steepest part of the rise has passed.
According to Mars Finance, on July 8th, Wall Street investment bank Bernstein released its monthly global memory tracking report, stating that DRAM contract prices continued to rise month-on-month in June, implying that the average price of traditional DRAM in Q2 2026 will increase by approximately 74% compared to Q1. Demand from servers and mobile devices remains the main driver, with Server DRAM prices expected to rise by approximately 60% to 67% in Q2, and Mobile DRAM by nearly 80%. The spot market also shows tight supply. PC DRAM spot prices rose 5.6% to 11.5% month-on-month in June, while Server DRAM rose 6.1% to 26.4%. The report stated that Server DDR5 performed particularly strongly, with spot prices significantly higher than contract prices, indicating that demand from AI and cloud service providers is still absorbing the new capacity shifted by suppliers to the server market. However, Bernstein also cautioned that the rate of price increases will slow significantly in Q3. TrendForce predicts that traditional DRAM price increases will slow to 13% to 18% in the third quarter, down from the steep increases in the second quarter. PC, mobile phone, and consumer electronics customers are reducing configurations or adjusting purchasing paces; while demand disruption is not yet fully apparent, the report believes it will eventually occur. The situation for NAND is more divergent. In June, NAND wafer spot prices actually fell by 3% to 4%, while wafer contract prices rose only slightly by 0.3% to 3.7%. However, strong price increases in mobile NAND and SSDs will still drive overall NAND contract prices up by about 60% in the second quarter. Bernstein stated that SSD and mobile storage price increases could reach 70% to 80%, offsetting the weakness on the wafer side. The key variable in this storage cycle remains AI. Cloud service providers and server customers continue to prioritize securing supply; some US CSPs have completed long-term agreement negotiations, while Chinese CSPs are still negotiating. Long-term agreements help smooth future price declines but may also limit the room for some suppliers to continue raising prices. In its investment conclusions, Bernstein maintains positive ratings for Samsung, SK Hynix, Micron, and SanDisk, while remaining cautious on Kioxia. The report believes that storage prices may remain strong until 2027, but will gradually normalize from the second half of 2027 to 2028 as long-term agreements take effect and new capacity comes online.