AI challenger from China causes a stir

05.02.202501 min

The Chinese company DeepSeek develops powerful artificial intelligence (AI) and could realign competition in the industry. While Western companies like OpenAI are investing billions in their models, DeepSeek takes a more cost-effective approach. Founded in July 2023 by Liang Wenfeng in Hangzhou, the company focuses on developing open-source language models.

The picture shows the user interface of a cell phone on which the AI app Deep Seek is open. It is structured like ChatGPT and other AI language models.
User interface of the DeepSeek app
Source: Pexels

Efficient AI models with low resource requirements

With its DeepSeek-R1 model, the start-up has developed an AI system that requires significantly fewer resources than comparable models from Western companies. The new AI model R1 is a language model designed for tasks that require logical thinking, mathematical problem-solving and real-time decision-making. It is based on the transformer architecture and has been trained with extensive amounts of data to be used in various applications such as text generation, code comprehension and translations. What makes DeepSeek-R1 special is its efficient use of hardware. While OpenAI uses state-of-the-art Nvidia chips for GPT-4, DeepSeek-R1 was trained on older Nvidia A100 GPUs, which significantly reduced development costs. With a budget of just $7 million, DeepSeek was able to create a model that is said to rival Western alternatives.

Rapid proliferation and international attention

After releasing its first free chatbot app on January 10, 2025, DeepSeek quickly rose to become the most downloaded app in the US iOS App Store, overtaking even ChatGPT. This development caused Nvidia’s stock price to drop by 17% as investors recognized the potential impact on the market.

Impact on the Western tech world

DeepSeek’s rise has sparked both admiration and concern in the Western tech world. On the one hand, the company’s efficiency and innovative strength is recognized. On the other hand, there are allegations that DeepSeek has trained its models using data from Western companies. OpenAI, for example, accused the start-up of using ChatGPT responses for training, which DeepSeek, however, denies.

AI Trends 2025? Find out what trends to expect here.

New dynamics in the global AI competition

DeepSeek shows that Chinese companies are increasingly developing competitive AI technologies. It remains to be seen whether the start-up can compete with OpenAI and other Western providers in the long term. However, technological competition is likely to intensify.

Jule Langheim studied media management at the Würzburg University of Technology. At eology she is part of the marketing team responsible for creating content and marketing the agency via social media channels.

Jule
Langheim
, Marketing Manager j.langheim@eology.de +49 9381 5829048

Contact
get in touch

  +49 9381 5829000