DeepSeek is here and China may have a rival for ChatGPT

The AI assistant has been developed by Chinese developers who shared their work, in an attempt to bypass US ban on AI chips.

Technology investors in the US and Europe woke up to a shock Monday morning, as Wall Street saw shares in US AI firms tumble, triggering sell-offs of more than $1 trillion. The reason: The arrival of DeepSeek, the Chinese AI assistant and chatbot that has taken the AI industry by storm and left rivals bamboozled since releasing its latest model last week.

On Monday, it was the top-rated free app on Apple’s App Store, overtaking US rivals such as ChatGPT and others, the BBC reported.

This matters because, according to its developers, the app is far more cost-effective, requiring far fewer advanced AI chips than some of its global rivals. And that claim, if true, not only undermines the US’s position as the global leader in AI, but also raises questions around the valuation of major tech firms such OpenAI, DeepMind and others.

It also puts the billions of dollars of investment in AI by firms in the US and Europe under the spotlight, leaving investors in doubt about the long-term prospects of their investment.

“We’re done following. It’s time to lead.”

Liang Wenfeng, CEO, Deepseek

The development also comes at a time when the US is trying everything it can to stop China from taking the lead in the global AI race, most notably by banning the export of advanced AI chips such as Nvidia to China.

Hancheng Cao, an assistant professor in information systems at Emory University says: “This could be a truly equalising breakthrough that is great for researchers and developers with limited resources, especially those from the Global South.”

The story of DeepSeek

The company was founded by Chinese information and electronic engineering expert Liang Wenfeng in 2023 in southern China. It is believed that he collected tens of thousands of Nvidia’s A100 chips before their export to China was banned. He later paired them with other cheaper chips that were still available on the Chinese market.

In 2024, Wenfeng told an audience at the China Academy: “We’re done following. It’s time to lead,” indicating his confidence at the time in the technology the firm was developing. Fast-forward to 2025, and the firm has launched the DeepSeek R1 model of its open source reasoning application, more commonly referred to as an AI assistant or a chatbot.

There are two key factors that set it apart from its global rivals, such as OpenAI’s ChatGPT. Both are crucial in deciding who leads the global AI race in the near future.

First, developers behind the DeepSeek R1 model claim that the whole project cost them around $6m. This is a fraction of the billions that US and other global tech firms are pouring into their AI projects.

It was possible because the model is reported to have required far fewer advanced AI chips to reach its current operational capacity, reducing the cost significantly.

The second factor that gives DeepSeek the upper hand is what experts have called “the engineering simplicity” of the R1 model. Its “chain of thought” approach enables it to “solve problems by processing queries step by step.”

Dimitris Papailiopoulos, principal researcher at Microsoft’s AI Frontiers research lab, explained: “DeepSeek aimed for accurate answers rather than detailing every logical step, significantly reducing computing time while maintaining a high level of effectiveness.”

How did markets react?

The Financial Times reported Monday morning that: “The tech-heavy Nasdaq Composite index slid 3.5 per cent shortly after the opening bell, while the broader S&P 500 fell about 2 per cent.”

The paper also claimed that Nvidia, the world’s largest AI chipmaker and one of the most valuable company globally, lost around $620 billion in market valuation, making it one of the worst victims of DeepSeek’s sudden rise to fame.

The shockwaves went beyond the US, with European tech giants such ASML also seeing an 8% drop in value. Other European firms such as as software giant SAP and Schneider Electric were also affected. European stocks recovered from the early morning shock and were heading back into positive territory around Monday afternoon.

The impact reaches beyond the traditional tech industry. Germany’s Siemens Energy – which supplies electrical hardware for AI infrastructure – felt a drop of 22% in market valuation. And there was bad news for the construction industry too. The FT reported that shares such as Spanish group ACS also dropped, down 6%, while Sweden’s Skanska fell 5%.

Experts at Goldman Sachs warned that, given the size and concentration of tech firms in the US index, there was always the risk of a negative spillover into the wider market.

Long-term implications

The arrival of DeepSeek on the global AI scene could have significant long-term implications beyond the temporary stock market fluctuations and investors’ fears.

The open rivalry between the US and China for global dominance in AI is at the heart of this. Beijing has seriously undermined Washington’s position as the undisputed champion in AI and advanced technology.

For many years, the US believed it could rein in China’s advances by imposing bans on key components, such as advanced AI chips used for developing large language models (LLMs). That policy seems to have backfired now.

Matt Sheehan, an AI researcher at the Carnegie Endowment for International Peace says: “The US export control has essentially backed Chinese companies into a corner where they have to be far more efficient with their limited computing resources.” 

Also, the open-source nature of DeepSeek’s entire project means the technology, and the details of all the technical work that went into developing it, will be available for everyone, and for free.

The arrival of DeepSeek on the open market, at a cost that embarrasses its US rivals, has created serious doubts among investors.

The firm has said on its website it has made that choice in order “to support the community.” But the move could also challenge the current hegemony of a few selected, mostly US-based, tech firms over the entire global AI landscape.

Lastly, and perhaps most importantly, it is about money. As discussed earlier, the arrival of DeepSeek on the open market, at a cost that embarrasses its US rivals, has created serious doubts among investors.

Only last week President Donald Trump announced the Stargate Initiative, a $500 billion privately funded project to develop AI infrastructure in the US. Key investors included OpenAI, SoftBank and Oracle.

The emergence of DeepSeek could jeopardise the future viability of such mega projects.