Singapore's largest money laundering case suspect Wang Shuiming may be extradited to China, with his accomplice being the mastermind behind the Hong Kong crypto platform scam AAX.

**June 19** Market analysts are broadly forecasting that SK Hynix’s Q2 2026 operating profit will land between 60 trillion and 70 trillion Korean won—eclipsing the prior quarterly record of 37.6 trillion won set in Q1. The April-released Q1 financial report showed revenue hit 52.58 trillion won, operating profit reached 37.61 trillion won, and operating margin stood at a blistering 72%—all marking quarterly highs. Red-hot demand and price hikes for High Bandwidth Memory (HBM) and high-end DRAM were the core drivers of this performance surge. Company management noted that structural memory tightness fueled by AI infrastructure will persist for at least three years. As Q2 kicks off, further upside to memory prices is lifting profit outlooks. Analysts project DRAM and NAND contract prices will climb 50% to over 70% quarter-over-quarter (QoQ). With high-margin product mix, each price hike creates outsized leverage for SK Hynix’s bottom line. The current market consensus pegs Q2 op

On June 19, the U.S. Dollar Index (DXY) extended its gains, climbing above 101 to reach its highest level since May 2025.

June 19: Per Whale Alert monitoring, an unidentified wallet address transferred 2,500 BTC to OKX at Beijing time today, valued at roughly $157 million based on current market prices.

June 19: Middle Eastern media outlets—including Saudi Arabia’s Hadassah, Turkey’s Anadolu Agency, and Iran’s Islamic Republic News Agency (IRNA)—report Israel launched fresh airstrikes on southern Lebanon’s Duweir, Harouf, and Na’bi Atiye areas on Saturday, causing casualties. (FXStreet)

Breaking: On June 19, the Swiss Foreign Ministry announced that the planned U.S.-Iran talks scheduled for Friday will not proceed as originally planned.

June 19 – Tang Jie, founder of SenseTime AI, shared in a post on X that after the official open-source release of its GLM-5.2 AI model, the system has delivered leading results across multiple major international authoritative assessments. In the Artificial Analysis Intelligence Index comprehensive evaluation, GLM-5.2 scored 51 points, placing it in the same performance bracket as Anthropic’s Claude Opus 4.8. It took 2nd place globally in the Code Arena front-end code generation adversarial test with an Elo rating of 1595, and claimed 1st spot in the DesignArena design-code integration scenario, scoring 1360 points. Overall, SenseTime’s GLM-5.2 remains a global leader in real-world use cases including front-end development, design generation, and software engineering. The model is steadily closing performance gaps with top models from OpenAI and Anthropic, and the team will continue pushing the limits of its capabilities. Earlier, when Elon Musk stated a leading Chinese large langua