封面新闻记者Yi Yili的计算能力时代，AI计算能力的重要性是自我言论的。 AI计算能力AI芯片的主要生命线一直是全球技术竞争的重点。 Noong Hunyo 6, si Pan Helin, isang kilalang ekonomista at miyembro ng Information and Communications Economic Expert Committee ng Ministry of Industry and Information Technology, ay nagsabi, "Noong nakaraan, dahil sa kakulangan ng mga lithography machine at EDA, ang China ay nasa isang passive na sitwasyon sa larangan ng chip. Gayunpaman, na may patuloy na layout ng mga kumpanya tulad ng Huawei, ang Gap sa AI Computing Power ChipsIT by 5-6 H20s can achieve H100 computing strength. There are many chips, which can increase electricity consumption and can cause loss of computing strength with bandwidth connection, but as long as adequate AI chips are stacked, the same effect of clumps of H100 and H200 chips can be达到的。t可以使用计算功率芯片的量来替换单卡的计算强度不足。中国人工智能发展中的真正瘟疫是计算强度兼容性。幸运的是，在年初DeepSeek发行之后，由于计算能力较低，国内AI芯片适应了DeepSeek。 NVIDIA在AI芯片中的主要竞争在哪里？潘·海林（Pan Helin）说，NVIDIA的芯片是由TSMC制成的，因此TheChip硬件不是Nvidia的优势。它的优势是软件，即CUDA生态系统。 NVIDIA AI芯片具有良好的兼容性，可以在无人机中产生图形，长生不老药或安装，并且可能是我的。在中国，Ascend采取了多申请兼容性途径，而其他人（例如Haiguang和Cambrian）采取了专门的芯片途径，该路线是不兼容的，只能用于AI培训。总的来说，目前的国内希望之星仍然是华为的上升，这似乎不仅仅是泰晤士报之一。