$Advanced Micro Devices (AMD.US)$ In general, A Mei's MI300 has made great efforts on memory, targeting the pain points of large models. They are also confident in comparing with Hopper, which is my Wei's H100. A Mei's current size and this effort are very sincere. But it must be known that Hopper is actually a semi-finished product (for large models). At that time, my Wei did not expect that large models would be so revolutionary, so you can see that the memory of H100 is relatively poor because it was not originally designed for betting on large models. Therefore, the 100% strength solution given by the next generation of my Wei for large models is really unimaginable. The AI cake has already been enlarged by various giants. From a risk perspective, all-in is not advisable. If you only invest in my Wei and A Mei, I think a position ratio of 6:4 or 7:3 is feasible. Of course, you should not only all-in on AI. The whole portfolio can be more diversified, for example, 50% in AI, and then invest in different AI symbols. However, everyone has different styles, and the key is risk tolerance. Congrats AMD! $NVIDIA (NVDA.US)$