The fact that a Chinese startup trained its AI model cheaply caused a stock price crash in the USA. For experts, however, this is a sign that artificial intelligence can be developed more quickly - including in the automotive industry.
A chatbot from China shocks Wall Street: Panicked investors have caused the market value of the chip company Nvidia to collapse by almost $600 billion in a single day. The trigger was the realisation that software with artificial intelligence (AI) could potentially be run with much less computing power - and thus with Nvidia chips - and that the technology can be trained much more effectively than previously supposed. The Chinese start-up DeepSeek claims to have trained its new AI model at a cost of less than $6 million, and on just a few stripped-down Nvidia chip systems. To what extent this information is accurate is currently still unclear.
Likewise, there are currently no publicly known collaborations between DeepSeek and automobile manufacturers. “The only public reference to the automotive industry is that TuanChe - a well-known online marketplace for automobiles in China - has decided to use DeepSeek’s application in the field of holographic AI,” Pedro Pacheco of management consultancy Gartner told automotiveIT.
“It is not yet certain who will win this race”
However, if DeepSeek’s statements prove to be correct, this would have great strategic significance for the automotive industry, says Pacheco. On the one hand, DeepSeek’s model is true open-source, which means that in principle, it can be adopted more easily by automotive manufacturers and technology companies that work with them. At the same time, the situation signals that there is now a real race for dominance in the field of AI between the USA and China.
“European car manufacturers cannot therefore afford to ignore Chinese technology providers, as it is not yet certain who will win this race,” Pacheco clarifies. Furthermore, similar to vehicle software, European OEMs that want to be leaders in AI would have to move the centre of their AI competencies to where the best AI talent and technology providers are likely to be - and that most likely means the US or China, the expert concludes.
Read more AI stories
- AI: Enhancing predictive maintenance in production lines
- Audi wants to define new standards for AI image processing
- The robots are here: How humanoids are changing auto production
- Data-driven decision-making for sustainable automotive production
What can the US achieve with better chips?
But what does all this mean for the future? American companies have recently outdone eachother with announcements about how much money they want to invest in AI infrastructure. ChatGPT inventor OpenAI and several partners alone have promised to invest $500 billion in datacentres in the coming years. At Facebook group Meta, founder Mark Zuckerberg has promised $60 billion this year alone.
But does DeepSeek now show that significantly less computing power will be needed?
US experts prefer to see it the other way around. The question is not whether DeepSeek can overtake the current market leaders in the USA, emphasized X. Eyeé from the AI consulting firm Malo Santo. It is much more about how quickly the Chinese research approaches can be implemented. “If DeepSeek can develop this with old hardware, what can we do with newer hardware?” she asked on the US broadcaster CNBC.
Who is behind Deepseek?
The company was founded in 2023 by hedge fund manager Liang Wenfeng - and is said to have secured a package of Nvidia chips. Deepseek relies on open source models where the source code is publicly accessible. The data is stored on servers in China. Information that is not acceptable to the Chinese government, such as the Tiananmen Square massacre in 1989, is withheld by the chatbot. The fact that Deepseek rose to number one in the US App Store for the iPhone is therefore particularly ironic in view of the tough crackdown on Tiktok.
With material from DPA
No comments yet