Tech stocks around the world were shaken throughout January 27 when news started breaking of China’s answer to ChatGPT – DeepSeek.

US tech giant Nvidia saw its value drop by more than a sixth, according to multiple reports following the rapid rise in popularity of the Chinese artificial intelligence (AI) app.

Launched last week, DeepSeek quickly became the most downloaded free app in the US. Developed at a fraction of the cost of its competitors, its sudden success led to sharp declines in the stock values of Nvidia and other AI-linked tech firms, including Microsoft and Google, during trading on January 27. However, claiming large-scale cyberattacks, DeepSeek announced the same day that it would temporarily be limiting registrations.

The chatbot is powered by the open-source DeepSeek-V3 model, which researchers claim was developed at a cost of approximately $6mn, much lower than the many billions spent by rivals although this is now a claim that has been met with scepticism by some in the AI sector claims the BBC. DeepSeek’s creators supposedly utilised existing technology and open-source software, which can be freely accessed and modified to help keep costs down.

Founded in 2023 by a graduate in information and electronic engineering, Liang Wenfeng in southeastern China, reports on the origins of DeepSeek are now suggesting Liang managed to amass a significant supply of Nvidia A100 chips before combining them with lower-end chips to develop his technology.

That one person could import Nvidia A100 chips, now banned for export to China, is already raising red flags, but this is a man already perhaps well-connected politically - having recently been seen at an industry meeting attended by Chinese premier Li Qiang

Similar red flags pop up in the form of reports and videos across the Internet showing the DeepSeek reaction to topics the ruling Chinese Communist Party might deem sensitive – the detention camps euphemistically called re-education centres filled with Uyghyrs in China’s northwestern Xinjiang region; also called East Turkestan, Taiwan’s status as an independent country, and the 1989 Tiananmen Square massacre.

Even questions on Winnie the Pooh are censored according to Forbes; the cuddly bear being a global 3D synonym of sorts for Chinese President Xi Jinping.

Tech journalists and YouTubers are already publicising a now common response to such topics when punched into DeepSeek of “Sorry, I’m not sure how to approach this type of question yet. Let’s chat about math, coding and logic problems instead!”

The emergence of DeepSeek comes at a time when the US is restricting exports of advanced chip technology to China, which has prompted Chinese developers to share resources and adopt innovative approaches to AI. These methods have resulted in AI models that require less computing power and are significantly cheaper to develop, potentially disrupting the global AI industry.

As a result, market analysts have noted that the arrival of a low-cost Chinese AI model has taken many by surprise, raising concerns about the profitability of firms that have heavily invested in more expensive AI infrastructure.

On the business front, some experts now suggest that DeepSeek could disrupt the investment landscape for the AI supply chain, while others caution that challenges faced by Chinese firms, such as access to advanced chips, could hinder their progress.

Add to this the fact that the man behind the software already appears to be mixing in political circles and is heading a brand that is openly censoring subjects that might upset Beijing, and DeepSeek is in for a rocky future.