In the past week, DeepSeek has caused great panic in the US investment community.
Even Trump urgently summoned Nvidia's Huang Renxun.
However, according to the feedback from Dan Ives, a well-known investor in the US technology field, after interviewing dozens of investment institutions and industry experts,
their investment strategy for AI infrastructure has not changed.
Nvidia's chips are still far from in short supply.
Let me use the Korean War as an analogy. In the early stages of the war, the Volunteer Army defeated the US Army with strategic and tactical innovations.
However, the US military found that the Volunteer Army's ammunition and logistical supplies could only support about a week, so they adopted a tactical innovation of retreating first and then counterattacking.
Another thing is that they did not reduce the use of firepower and ammunition, and even greatly increased it.
This is the origin of the famous "Van Fleet ammunition".
The current strategic goal of artificial intelligence is to see who can take the high ground of ASI (super artificial intelligence) first, and a large number of robots that work effectively in the physical world.
If China and the United States tie in terms of algorithms,
then large-scale high-performance chip computing power will still be the winning point.
In the long run, the demand for computing power for the application of artificial intelligence in the world is far from reaching its peak.
Make an analogy with electricity and bandwidth.
In addition, DeepSeek is a breakthrough result obtained after distilling a large amount of training data from OpenAI.
"Standing on the shoulders of giants" (Liang Wenfeng's words)
Then what if there are no giants for you to stand on?