ETF Volatility | Southern 2X Long NVIDIA rises over 6% as NVIDIA stock price reaches a new high; its platform adapts to the DeepSeek-V4 AI model
Southern 2X Long NVIDIA rose over 6%, as of the time of writing, up 6.32%, reported at HKD 183.35, with a trading volume of HKD 1.8865 million. On the news front, technology stocks in the U.S. such as Alphabet, Microsoft, Amazon, and Meta will announce their earnings this Thursday (April 30), and market attention on AI-related performance continues to rise. As of the close on April 27, NVIDIA closed up 4%, with its stock price breaking new highs and a total market value exceeding USD 5.25 trillion. It is also worth mentioning that recently, NVIDIA published a blog post announcing that its NVIDIA Blackwell platform has been adapted for the DeepSeek-V4-Pro and DeepSeek-V4-Flash models. Test data shows that DeepSeek-V4-Pro achieves out-of-the-box performance exceeding 150 tokens/sec/user on the NVIDIA GB200NVL72, and with the vLLM's Day0 recipe, developers can quickly deploy on the Blackwell B300. With deep optimizations of Dynamo, NVFP4, and CUDA kernels, expected performance will further improve
According to Zhitong Finance APP, the Southern 2X Long NVIDIA (07788) has risen over 6%, currently up 6.32% at HKD 183.35, with a trading volume of HKD 1.8865 million.
On the news front, major U.S. tech stocks such as Alphabet, Microsoft, Amazon, and Meta are set to announce their earnings this Thursday (April 30), with the market's focus on AI-related performance continuing to heat up. As of the close on April 27, NVIDIA's stock rose 4%, breaking new highs, with a total market capitalization exceeding USD 5.25 trillion.
It is also worth mentioning that recently, NVIDIA published a blog post announcing that its NVIDIA Blackwell platform has been adapted for the DeepSeek-V4-Pro and DeepSeek-V4-Flash models. Test data shows that DeepSeek-V4-Pro achieves out-of-the-box performance exceeding 150 tokens/sec/user on the NVIDIA GB200NVL72, and developers can quickly deploy on the Blackwell B300 using vLLM's Day0 recipe. With deep optimizations of Dynamo, NVFP4, and CUDA kernels, further performance improvements are expected
