$NVIDIA(NVDA.US) returns to the most basic simplified formula: AI = data + algorithm + computing power.

In the open-source world, data doesn't differ much among companies, and algorithms are similar. Previously, major giants mainly competed in computing power (GPU).

Now, with the emergence of DeepSeek, there's a significant wave of competition in algorithms, causing stock price fluctuations for giants focused on computing power.

DeepSeek reduces computing power demand to 1/10 of traditional methods. This means:

1) Short-term market reliance on top-tier GPUs drops significantly.

2) Short-term hardware deployment scales down substantially.

3) Short-term overall computing power costs decrease sharply.

For the semiconductor manufacturing industry, which heavily depends on hardware iteration cycles, this shift in technological approach—from "stacking GPUs" to "optimizing algorithms"—is significant.

Short-term impacts are felt by NVIDIA and AMD due to reduced computing power demand.

However, DeepSeek is open-source, making its algorithms easily replicable by others. Additionally, DeepSeek isn't a multimodal model, so it has usage limitations.

It's hard to imagine how terrifying the models trained by GPT or Claude would be with the same algorithm and equivalent computing power.

Moreover, there are rumors that DeepSeek uses restricted H100 GPUs, meaning no matter how advanced the algorithm is, GPUs are still required.

For example, to make high-speed trains faster, there are two approaches: better railways and trains (computing power) or more efficient stations and crew systems (algorithms). When the latter becomes similar, the competition still revolves around computing power.

So, the conclusion is: short-term bearish for NVIDIA, but long-term, computing power remains a fundamental demand and won't change.

As for how much the stock price will drop, I don't know. What I do know is that if it drops too much, it might present an opportunity (but risk control is essential).