China's catch -up in artificial intelligence: the case Deepseek and the strategic data use

Published on: May 18, 2025 / update from: May 18, 2025 - Author: Konrad Wolfenstein

China's catch -up in artificial intelligence: The case Deepseek and the strategic data use - Image: Xpert.digital

The new AI world power: China pulls the same with the USA

With the sudden success of the Chinese AI company Deepseek in early 2025, the world clearly showed that China not only catches up for artificial intelligence, but is already on par with the leading US companies in some areas. This development is not a coincidence, but the result of a targeted strategy based on massive data use and efficient scaling - with special effects on the health sector and industrial production.

Suitable for:

Artificial intelligence: BMW relies on China-Ki with Deepseek

Deepseek: From the unknown start-up to the AI forwarder

The company Deepseek, founded by Liang Wenfeng in July 2023, moved into the international spotlight in early 2025 with the publication of his language model Deepseek-R1. The model exceeded American AI models such as Openais GPT-4 and O1 in several benchmark tests, especially in the areas of mathematics and readering. The Chinese company app even took first place in the download charts in the US App Store-before chatt.

Particularly noteworthy: Deepseek achieved this performance with significantly lower resources. The company states that it has trained its model for only about $ 6 million-a fraction of the estimated $ 100 million that Openai spent on GPT-4. The computing effort is also said to have only about a tenth compared to comparable models such as Metas Llama 3.1.

Efficient AI development despite trade restrictions

This efficiency lead is all the more remarkable because Chinese companies have no access to the most powerful AI chips due to US export restrictions. Deepseek therefore developed alternative approaches such as “Mixture of Experts” (MOE), in which specialized sub-models are activated depending on the request-an architecture that the European AI model Mistral also uses. Instead of relying on raw computing power, Chinese developers optimize their software for weaker but exportable chips.

This forced situation ultimately turned out to be an innovation driver: "Unlike many Chinese AI companies that depend heavily on progressive hardware, Deepseek has focused on maximizing software-powered resource optimization," explains Marina Zhang, professor at the University of Technology Sydney.

China's strategic approach in AI development

Deepseek's success is not an isolated case, but part of a comprehensive national strategy. As early as 2017, China published its “New Generation Artificial Intelligence Development Plan”, which defines clear goals: by 2025, the country is to open up the global leadership in the AI in key areas such as autonomous driving, robotics and AI-based health care to the top of the world by 2030.

Made in China 2025 and the AI priority

The AI strategy is embedded in the larger plan “Made in China 2025”, which aims to lead China into the higher value chain. This plan, initiated by the Chinese Prime Minister Li Keqiang, stipulates to increase the domestic proportion of core components to 70% by 2025, with special focus on high-tech areas.

The current five-year plan (2021-2025) and the new “AI+” initiative underline the central importance of the AI for China's economic future. The initiative aims at the connection of digital technologies with China's industrial strength and its huge internal market. For example, Beijing has published an action plan that comprises eight measures to support IT software companies that apply AI large models in various industries-with funding of up to 30 million yuan (around 3.8 million euros) per company.

The data advantage: key to China's AI catching up

A central advantage of the Chinese AI strategy lies in access to huge amounts of data. "While western countries work with high hurdles in data protection and ethical guidelines, China acts within a state -controlled framework that explains the use of sensitive data on the economic and security policy priority," analyzes Renata Thiébaut in the Handelsblatt.

Targeted data use in medicine

This competitive advantage is particularly evident in the field of medicine and genetics. The “Human Genetic Resources Regulation” updated in 2023 allows targeted use of genetic and health -related data under state control. These framework conditions enable faster clinical studies and more efficient active ingredients.

In Chinese health care, AI applications such as Deepseek are already used to optimize processes. At the Second Affiliated Hospital of the Medical University of Fujian, Deepseek, for example, supports the electronic patient files by automatically summary of patient data and help in medical documentation.

This development also manifests itself in specific projects such as the “Ai Agent Hospital” opened in Beijing in 2025. This hospital, which was designed by the renowned Tsinghua University, integrates KI comprehensively into the design and operation of the hospital system. In particular, it should counteract the shortage of doctors in rural and sub -provisioned areas and support medical decision -making.

Structural factors promote AI use in healthcare

The Chinese market for medical AI benefits from structural circumstances: "In China there is a lack of doctors per capita, but an abundance of IT engineers. Medical resources are concentrated in the cities, while they are scarce in rural areas". Both problems can be alleviated by AI supported telemedicine.

In addition, the strong networking of public hospitals, which make up more than 80% of inpatient and outpatient treatments, enables the training of AI models with provincial second data sets. On the other hand, US researchers have to work with data from individual private clinics that operate under stricter data protection rules.

Suitable for:

Economic Turbo Deepseek: China's new AI hope as an economic engine?

Industrial application: AI as an engine for production

In addition to healthcare, China drives AI integration ahead, especially in industrial production. The “AI+” strategy explicitly aims to network and upgrade the entire Chinese industry. AI is supposed to produce goods and steer their currents, which can significantly increase productivity.

In contrast to many western companies, the Chinese AI companies founded in 2023 do not primarily rely on basic research, but on implementation competence: “While the focus is still on basic research in the USA, China scores with implementation skills”. This pragmatic approach enables faster integration of AI solutions into existing production processes.

How China's AI masters a tightrope walk between progress and control

Despite the impressive progress, China's AI development faces challenges. The international debate about Deepseek illustrates the tensions between technological progress and political control. Deepseek's AI refuses to answer politically delicate questions, such as the role of Chinese President Xi Jinping or the Tiananm Proteste 1989.

This censorship raises questions about the relationship between private companies and state control. Although Deepseek is described as “relatively far from the Chinese government”, the presence of the CEO Liang Wenfeng shows the close integration of business and politics in China at a meeting with China's Prime Minister Li Qiang.

At the same time, the recent successes of Chinese AI companies are celebrated as national achievements in the state media. This development has geopolitical dimensions: "A globally used AI would be a powerful tool to enforce geopolitical interests". China is repeatedly criticized internationally because of disinformation campaigns and hidden influence.

Systemic competitive advantage through strategic orientation

China's rapid rise in AI development is based on a systemic approach that combines state planning, targeted data use and pragmatic implementation. The success of companies like Deepseek is not a coincidence, but the result of this strategy, which is particularly used in healthcare and industrial production.

The Chinese AI development benefits from scale effects of the large domestic market, strategically oriented grants and more flexible use of data. Paradoxically, US export restrictions on AI chips have increased the speed of innovation by forcing Chinese companies to develop more efficient architectures.

This development presents Western technology companies and governments with new challenges. The global race for AI leadership has received a new dynamic, with China with its targeted strategy proving that technological innovation does not depend solely on financial strength, but also on systemic advantages and strategic planning.

Suitable for: