- European businesses are increasingly considering the use of Chinese open-weight AI models due to their low costs and competitive performance.
- According to Arthur D. Little, operating Chinese AI models on infrastructure located in Europe can offer higher operational control compared to using proprietary US AI services.
- Expert Volker Pfirsching argues that technological sovereignty should not be judged solely based on the provider’s nationality.
- Open-weight models allow businesses to download, modify, and deploy them on internal servers, ensuring that sensitive data remains within European control.
- However, using Chinese foundation models could also create new dependencies in the supply chain, contradicting the European Union’s goal of technological autonomy.
- According to Pfirsching, many businesses realize that using the most powerful AI model for every task is not cost-effective and are opting for models suitable for each specific workload.
- Last month, China’s Xunce Technology partnered with Lutech to bring the TokenOS platform to Europe and build an “AI token factory” that meets the region’s data regulations.
- According to the Financial Times, Siemens has tested Alibaba’s DeepSeek and Qwen models in several applications.
- Businesses using Chinese AI models must still comply with the General Data Protection Regulation (GDPR), including strict requirements when data can be accessed from China.
- Experts emphasize that self-hosting models in Europe does not eliminate the need for cybersecurity assessments, software supply chain control, and compliance with the EU AI Act.
- Many European businesses are heading toward a strategy of combining US, European, and Chinese AI models to reduce dependency on a single provider.
- According to Arthur D. Little, Chinese AI models will become an important part of that portfolio due to their increasingly competitive performance and cost-efficiency.
📌 Europe is approaching AI with a more pragmatic mindset rather than one based purely on geopolitics. Chinese open-weight AI models are seen as attractive because they can be deployed on internal infrastructure, allowing businesses to maintain data control and optimize costs. However, adoption must still meet strict requirements regarding security, GDPR, and the EU AI Act, while many businesses choose a provider diversification strategy to avoid dependency on any single AI ecosystem.
