DeepSeek Send Panic Shivers Down the US AI Industry

A new Chinese AI model has taken the world by storm. Chinese AI startup DeepSeek has just released the latest version of its open-source AI model – DeepSeek V3, and it’s better than the latest ChatGPT 4.0, Llama3, and Claude AI models. In fact, in some areas, it even surpasses its rivals.

To add a cherry to the top, its training and development cost is just a fraction of that of its rivals. Where giants like OpenAI and Meta have pumped billions of dollars into data centers and AI training, DeepSeek has built a model from scratch for just $5.5M. 

The US stock market bore the brunt of this new breakthrough, wiping off $1 trillion for investors in a single day. Nvidia was the most affected stock – it fell almost 17% in a day. Nvidia investors saw $593B vaporize into thin air – the single biggest market cap loss for a company on Wall Street. 

Nvidia’s whole business is centered around selling massively expensive GPUs at 80-90% margins. In contrast, DeepSeek built the whole model on around 2,000 Nvidia H800 GPUs. 

DeepSeek has rocked the belief that the AI industry is a capital-intensive one and only big corporation giants can participate in it. 

As per reports, DeepSeek uses efficient computing techniques such as a “mixture of experts” approach to lower the cost. Now that the industry knows similar results can be achieved with a fraction of resources, the whole AI chip industry business may falter.

The only downside might be that the model is trained on data up to October 2023. But this is only for the web version. The app seems to have access to the latest information. 

What Does This Mean For US AI Giants?

DeepSeek stepping foot into the industry has sent the rest of the players into a frenzy. As per an inside source, the generative artificial intelligence division of Meta is in “panic mode”.

Engineers are moving frantically to dissect DeepSeek and copy anything and everything we can from it. I’m not even exaggerating. – Meta employee

The biggest worry of the management right now is how they will justify the massive cost of their tools and the massive salary that their “leaders” are drawing when China has managed to create an equally competent tool at such a low price. 

Even President Donald Trump praised DeepSeek in a speech at a Republican retreat, applauding its technological breakthrough and saying it should be a wake-up call for American companies.

OpenAI’s CEO Sam Altman seemed to be a little more calm in his response. He took to X and said that it was quite “invigorating” to have new competitors in the market. He also applauded the company for delivering that kind of service for such a low cost and in the end, promised to speed up some of OpenAI’s own releases.

It might seem that DeepSeek is the next big thing. And it might be so, considering that it has already become the No.1 downloaded free app on Apple’s app store. Its open-source model

makes its codes and technical details open to all. This might also speed up progress through collaboration. 

But there are some hurdles in the way. For instance, so far the tool censors all topics that are not favored by the Chinese government. However, the question is how viable this approach will be if it tries to make it big in the global market. 

Plus, China is not on good terms with many of its neighbors. We have all seen how it went down for TikTok. So in this environment, will DeepSeek manage to survive or will it sink after this banging launch?

Vlad is Tech Report’s Executive Editor. With over a decade of experience in tech content, he’s passionate about computer hardware, an advocate of online privacy, and strongly believes in the open-source, scarce-money nature of cryptocurrency.
When he’s not working, he’s traveling with his partner and their cat, learning Python, or reading good books. He never owned a PC he did not build.

View all articles by Vlad Melnic