Summary of Key Points
The competition between China and the United States in AI in emerging markets (such as Africa) has shifted from focusing on model performance to emphasizing infrastructure, financing capabilities, and regulatory adaptability. China relies on the Digital Silk Road to offer integrated solutions that include infrastructure, financing, and open-source models, while the U.S. emphasizes hardware control and systematic output through its export initiatives. Models are no longer considered a competitive advantage; rather, success depends on national-level resource coordination, diplomatic strategies, and the ability to implement infrastructure.
1. Models Are No Longer “Exclusive Secrets” – They Have Become “Commodities”
In the past, AI competition was about which model ran faster or made more accurate predictions. However, this has changed. Both Chinese and American models are now “good enough” for most use cases. For example, China’s Qwen model has been downloaded more times than Meta’s Llama, accounting for 45% of global downloads, not because of its superior performance, but because it is open-source (free or low-cost) and highly adaptable, making it the default choice for developers. The U.S. AI export strategy does not even require high model performance; instead, it focuses on hardware control and access rights – indicating that the difference in model capabilities is minimal, and there’s no need to compete solely on this aspect.
2. The U.S. AI Export Strategy: Emphasizing Hardware Control and Systematic Packaging
The U.S. offers a comprehensive set of AI solutions, from hardware to cloud services to industry applications. A key requirement is that at least 51% of the hardware must be American-made, and the model must be controlled by a U.S. company, which is responsible for its maintenance. This reflects the U.S. belief that while model performance is not crucial, it is essential to control both the hardware and the underlying systems.
3. China’s Approach: Using the “Digital Silk Road” to Offer Integrated Solutions
China follows an upgraded version of Huawei’s African strategy, combining digital infrastructure (communications, payments, data centers), national financing guarantees, and AI technology. For instance, Huawei Cloud provides integrated “cloud + AI” solutions in countries like Saudi Arabia and Thailand. Chinese smartphone brands and digital payment systems have already established a foundation in Africa, making it easy to integrate AI services. Domestic open-source models, such as DeepSeek and Qwen, focus on affordability, adapting to the budgetary constraints of emerging markets. They do not aim for extreme performance but ensure compatibility with limited resources.
4. Infrastructure and Regulation Are Critical Determinants
The biggest barriers to AI adoption in emerging markets are infrastructure and policies:
- The $1 billion data center project in Kenya failed due to a lack of long-term energy guarantees and financial support from the government.
- South Africa retracted its AI policy because the references used in the policy were generated by AI, indicating inadequate regulatory capabilities.
- NVIDIA’s success with Kasawa was due to Kasawa’s extensive fiber optic network and renewable energy infrastructure – these are the real determinants of success.
5. Can the Huawei Model Be Replicated? Similar, but with Challenges
Huawei’s approach of combining infrastructure and financing to enter African markets can be replicated. However, there are differences: core AI components (chips, cloud architecture) are more difficult to replace and are subject to stricter export controls. While Huawei’s base stations can be replaced with domestic alternatives, replacing AI chips is more challenging. Therefore, while the “financing + rapid implementation” aspect of Huawei’s model can be copied, achieving autonomy in core technologies is essential.
In Conclusion
AI competition is no longer about which model is smarter; it’s about who can provide the necessary infrastructure, financing, and policies to enable AI adoption in emerging markets. China has an advantage in this regard, having already established a foundation in infrastructure and financing. However, it needs to overcome challenges in core technologies. The U.S. excels in hardware control, but the cost of its systematic approach is higher. If there are breakthroughs in model technology, the focus may shift, but for now, the “foundation” (infrastructure) is more critical than the “intelligence” (models).