Short-term negative, long-term positive!

The day before yesterday, it was timely reminded that China's AI model Deepseek might have an impact on Nvidia and the US stock market, and now it seems to have caused a huge impact.
I just saw Jiang Zhuoer's statement that "Deepseek is beneficial for Nvidia." I looked up some information and would like to refute this:
Jiang Zhuoer's argument that "Deepseek only optimized computing cost and did not produce new breakthroughs" is incorrect. In fact, Deepseek not only excels in significantly reducing the cost of training models and enhancing inference capabilities but also surpasses traditional models (ChatGPT-4) which require dozens of H100-level GPUs (costing over a million dollars) for complete deployment. Deepseek has broken through the scale limitations of large model deployment through technological innovation, allowing R1 to run smoothly on a single RTX 4090 (24GB memory) with an inference speed of 15-20 tokens/second. This is quite frightening for Nvidia. Their current profits heavily rely on high-performance GPUs, and if everyone can use ordinary gaming GPUs to do AI, then that would be a big problem.
In any case, the capital market has given an answer~ choosing to be bearish on Nvidia. It is recommended to observe more for now and let the bullets fly a bit longer.