On August 7, SK Hynix announced a dividend of 375 won per share and stated that it will reveal its shareholder return policy in the third quarter, while also considering additional shareholder return measures.

The KOSPI index in South Korea fell over 5% this week, marking its seventh consecutive week of decline, the longest losing streak since December 2022.

On August 7, the Japanese Government Pension Investment Fund announced a record quarterly earnings of 24.1 trillion yen.

On August 7, Zhongji Xuchuang experienced a short-term drop of 4.5%.

On August 7, it was announced that the Korea Exchange will officially start after-hours trading for exchange-traded funds (ETFs) on September 14. The Korea Exchange aims to compete with alternative trading systems like Nextrade and all-weather cryptocurrency exchanges. In response, asset management institutions have warned that the lack of real-time net asset value estimates could widen the price deviation of ETFs; the industry had previously called for the Korea Exchange to postpone the launch of this service. Individual stock leveraged ETFs will not be included in the after-hours trading scope.

On August 7, news emerged that, according to the latest monitoring data from Arkham, Morgan Stanley has made another low-cost purchase, investing $7.21 million through its spot Bitcoin exchange-traded fund MSBT to acquire approximately 100.3 BTC. As of now, its total Bitcoin holdings have surpassed 6,300 for the first time, reaching 6,331, with a value exceeding $406 million.

On August 7, the Financial Times reported, citing three informed sources, that ByteDance has begun pre-training a large model with up to 100 trillion parameters. The project is still in its early stages, and the final scale has yet to be determined. If trained to the upper limit, it would exceed KimiK3 by more than three times, making it the largest model known among Chinese teams in terms of parameter scale. ByteDance has directly pushed the model size close to Anthropic's flagship. Although Anthropic has never disclosed the parameter count for Mythos5, the Financial Times cited industry estimates suggesting that Mythos5 has around 80 trillion parameters, while Fable5 has about 50 trillion. If ByteDance's new model achieves 100 trillion parameters, it would at least match the scale of the leading closed-source models in the U.S. This model is currently in pre-training (training the model's foundational capabilities using massive amounts of data). The Financial Times noted that this phase typically takes 3 to 6 months, followed by further post-training before a successful release. More parameters do not necessarily equate to stronger capabilities; the final performance also depends on architecture, training data, and training methods. Zhang Yiming recently made it clear internally at Seed that he opposes distilling competitor models, urging the team to accept short-term setbacks and strive to reach the world's top tier through self-research. This new model, with a maximum of 100 trillion parameters, represents ByteDance's most aggressive bet on scale to date.

On August 7, it was reported by informed sources that SK Hynix has accumulated purchases of approximately 100 trillion to 400 trillion Korean won (about $7 billion to $28 billion) in Korean bond assets this year, with high-end estimates including commercial paper. As of the end of the second quarter of this year, SK Hynix's cash and cash equivalents reached 88 trillion won, an increase of nearly 62% from the previous quarter. The sources indicated that SK Hynix began to expand its bond investments in April of this year, with individual order sizes typically ranging from 100 billion to 300 billion won, primarily focusing on investment-grade bonds rated AA or higher, with maturities generally not exceeding three years. (Bloomberg)

On August 7, it was reported that the UAE is considering an investment of 1 trillion yen (approximately 6.3 billion USD) to participate in the construction of an AI data center project in Japan. The Abu Dhabi sovereign wealth fund Mubadala Investment is expected to lead the investment and may collaborate with other overseas and local Japanese investors. Once completed, the project is anticipated to become Japan's largest AI data center, with total investments (including supporting enterprises and surrounding infrastructure) expected to reach up to 2 trillion yen (approximately 12.6 billion USD). It is reported that the project will deploy NVIDIA AI servers. The Japanese government has already prioritized data centers and cloud computing as key growth strategies, aiming to attract 32.7 trillion yen in public and private investments by the fiscal year 2035. In recent years, companies such as TSMC, Tower Semiconductor, and Micron have expanded their investments in Japan, while Japanese data center operator NTT Data plans to invest at least 9 billion USD to expand computing infrastructure by 2033. (Bloomberg)