After DeepSeek teamed up with Huawei Ascend chips and globally released the DeepSeek V4 version, Huang Renxun gave Nvidia a dead order! What is Huang Renxun really afraid of? On the early morning of April 24th, after OpenAI released GPT 5.5, Nvidia CEO Huang Renxun quickly did three things: sending a congratulatory letter to Altman, sending a mass email to about 30000 employees, and demanding the mandatory use of GPT 5.5 based Codex throughout the company. The most noteworthy thing is not the "congratulations" themselves, but the phrase "must use" in the email - engineering, legal, finance, and HR are all included, indicating that this is not a technical team pilot, but rather a direct push of AI tools into daily processes. The most notable feature of DeepSeek V4 this time is the migration of code from the CUDA ecosystem to the Huawei CANN ecosystem. Previously, DeepSeek used Nvidia chips for training and reasoning AI models, but now all have switched to Huawei's Ascend chips. From hardware to software, DeepSeek V4 is definitely a domestically produced product in China. What is CUDA? It is a software moat built by NVIDIA over nearly twenty years, and almost all AI developers around the world write programs and train models on it, just like how we are all used to Windows computers and Android systems. Now, DeepSeek V4 has teamed up with Huawei to dismantle what Huang Renxun and his team consider to be Nvidia's strongest moat! The 'enemy' has already begun to attack the city and plunder the land. Can Huang Renxun not panic? The scariest thing is that DeepSeeker V4 is an open-source AI model that countries and individuals around the world can use and transform into their own AI models. The most gratifying thing is that DeepSeek V4's cost-effectiveness is directly maximized, with the best overall cost-effectiveness! For many countries and individuals around the world, DeepSeeker V4 is really fragrant and tempting. To the AI industry in the United States, it is a sharp sword aimed at American closed source AI models and AI chips! In the past, AI models focused on computing power and chips. Nowadays, we not only need to compete for computing power, chips, and industrial chains, but also for ecology. Going forward, it may be up to who defines the game rules of AI. DeepSeek upgrade is about charging forward and conquering cities and territories. Huang Renxun instructed his employees to fully utilize OpenAI, which is using manual computing power to better sell chips and defend Nvidia's moat! Previously, people thought that AI competition was just a game for American giants. Now, Chinese players are sitting at the table and starting to play with American wrench wrists. No wonder some netizens commented: Since Huang Renxun requires employees to use OpenAI, shouldn't Chinese people use DeepSeek more? Please provide some support. What do you guys think?