The Hangzhou start-up DeepSeek Artificial Intelligence shook up the AI world on January 20 by releasing freely usable large language models (LLMs) presented as being able to rival those of major American companies (such as OpenAI and Google Gemini) in terms of performance. All this at a lower cost in terms of the IT resources required to deliver results to user queries.

It's enough to panic financial markets and investors in the AI sector and to raise questions about the resources needed to innovate, at a time when US President Donald Trump has just announced colossal investments.

Users can try out LLMs released by DeepSeek in a number of ways. First, through the free chat and query interface – a conversational agent very similar to the many AI-automated conversation applications, starting with OpenAI's famous ChatGPT. DeepSeek offers this service on its website and on a mobile application for smartphones. The French are currently downloading it in large numbers – on Tuesday, January 28, it was the seventh most downloaded application on Android in France, and the first on iOS.

You have 84.62% of this article left to read. The rest is for subscribers only.