DeepSeek is a Chinese artificial intelligence company that works to develop large language models. In early 2025, this DeepSeek AI startup started raising quickly and started getting a lot of DeepSeek AI hype. This hype caused a big sell-off in the Nvidia shares, wiping out hundreds of billions in value from the stock. On 27 January, Deepseek AI released their chatbot called R1, and in a very short time, it became one of the most downloaded free apps in the US. They even beat ChatGPT at it. After some time, DeepSeek chatbot became controversial because Nvidia’s stock fell by 17-18% in just one day. Many experts call this moment the Sputnik moment for AI.
The origins of Chinese AI startup DeepSeek
The Deepseek is built by the High-Flyer. The R1 model of it, which got a lot of hype, mainly gets trained using Nvidia A100 chips, and this all got bought before the US bans the expert and smart techniques, such as “mixture of experts,” that cut costs. The origins of Chinese AI startup DeepSeek are not new, and the people who use AI tools often or get in touch with AI news, even a little bit, surely have heard about this tool. DeepSeek also became one of the most used tools, but now it seems the company is going to face serious issues after using social media bot networks in large amounts.
How Bots Boosted the DeepSeek AI Hype
The investigation resulted in the discovery that all the hype was fake and Smarty created. More than 3000 accounts, which were created purposely to do the fake hype, were found by EVAI bot analysis. Those accounts were doing the paid promotion to create all the hype, and they posted photos in favour of the Chinese AI startup DeepSeek. Most of the posts were repeated by similar accounts. The company named Cyabra also confirmed the fake DeepSeek AI hype and said 15% of online buzz was from the fake account. It may be driven by politics and money.
Nvidia Fights Back After Selloff
Jensen Huang, Nvidia CEO, said that the market has overrated DeepSeek AI. He said that powerful chips are still essential for AI development, even after the training. The CEO said the whole DeepSeek AI hype about R1 is exciting, but we can’t forget the fact that Nvidia remains central to AI progress. This whole statement by him helped Nvidia a lot; they almost recovered all the losses they faced because of the DeepSeek chatbot controversy. At the same time, an investigation by the US official agencies also got started to check if DeepSeek got Nvidia chips through Singapore to be on the safe side from export bans.
DeepSeek Faces Tough Questions
After the crash, all tech stocks bounced back. Companies like Microsoft, Google, Meta, and Amazon helped Nvidia regain $1 trillion, and this all happened within two months after the fake DeepSeek AI hype got exposed. The regularities also started keeping an eye on Deepseek and how they are working, so they can’t do a similar act in the future.
This fake hype can become the reason for someone else who is genuine and actually beneficial, and the biggest example of it is Nvidia. Some also claim that the idea of building an R1 model with only $5-6 million seems fake; there are huge chances that they copied the data from OpenAI. Some also started raising questions about its privacy policy and about the controls of its sensitive political topics. Now, Nvidia is getting back on track, and DeepSeek has lost the trust of its many loyal users.
