The current problem with AI is no longer that whose large model is easier to use or more powerful, but that many companies only find out that it is too expensive after applying it on a large scale. Even American technology giants like Uber and Microsoft are trying to reduce costs. Ordinary people are in the same predicament. There are many people who run AI intensively every day and use hundreds of millions or even billions of tokens every day, which all costs money.At NVIDIA's internal meeting, some employees also asked questions, worried that tokens were just wasted, and only superficial efforts were made without actually improving productivity.

NVIDIA CEO Jen-Hsun Huang responded that in the early stages of any technology, people are often not familiar enough to use it freely, so don’t pursue perfection from the beginning. You need to take the first step to embrace new technologies.
As for the cost issue, he said it's okay to waste a little money, but the key is not to waste time.
Regarding the issue of burning money on Tokens, Huang Renxun’s attitude is very supportive of everyone’s extensive use of AI. He previously said that an engineer with an annual salary of US$500,000,If the token fee is less than $250,000, he will be very worried and will explode if he only uses $5,000.
Huang Renxun’s point of view is that he would rather give them half of their salary than let employees use AI to increase efficiency by 10 times.
However, the background behind Huang Renxun's remarks is that the company can reimburse employees for token expenses, and employees do not have to pay for them themselves, at least for most expenses. However, the current situation is that with the increase in usage, and token usage has become an indicator of employee assessment KPI, burning money is not a small problem.
Even Jen-Hsun Huang, who is worth a trillion dollars, only said when he answered the question this time that it was okay to waste a little bit, and he did not dare to say that he would use it casually. After all, American companies mainly use AI from OpenAI and Anthropic.The Token quotation is more than 50 times that of DeepSeek, and the money burning speed is not at the same level.