Chinese AI Expeditions over the West Ocean

We are making this article available for everyone. If you want to support Turing Post please →

these tech giants have centered their overseas strategies on Southeast Asia and the Middle East, aligning with China’s Belt and Road Initiative and the Digital Silk Road policy. Amid rising geopolitical tensions, choosing regions where Chinese is commonly spoken, such as Southeast Asia, or emerging markets like the Middle East and long-time allies like Africa, seems a more strategic choice.

ByteDance, referred to as "App factory," has chosen to focus on familiar Western consumer-focused (to-C) markets, launching 11 overseas applications in just seven months. CapCut, launched in 2020, released its paid version CapCut Pro in 2022, then integrated AI features in the beginning of 2024, now it became one of the world’s most popular apps by 2024, with over 300 million monthly active users. According to Sensor Tower, by July 2024, CapCut had generated $125 million in cumulative revenue from mobile applications.

Then comes the Chinese AI unicorn startups with a different strategy, by adopting a model + application approach. Facing high costs for training models, some have begun to shift focus from updating foundational models to more profitable application and scenario exploration. For startups reliant on funding, expanding overseas has become a necessity amid intense domestic competition. Many early-stage companies have chosen Western to-C markets, launching productivity, creative, and companion apps based on their respective models. For example, among the “Big Six”,  apps “Talkie” from MiniMax and “PopAI” from 01.AI (the profile) which has gained millions of users and achieved a level of profitability.

Why Expand Overseas?

Domestic Market Competition

China has the world's largest number of internet users and a vast pool of technical developers, and no one wants to be left behind in the AI boom. As more companies flood the space, AI technology has developed rapidly, but the growth of applications and use cases has been slower. Both industry giants and startups face growth stagnation and profit pressure.

Between October 2023 and September 2024, China released 238 LLMs. After more than a year of fierce competition, they entered a phase of consolidation. The pressure built up in May 2024 during a price war, triggered by DeepSeek (Read Inside DeepSeek Models), an AI startup, which introduced architectural innovations that significantly reduced model inference costs. Following the announcement, major players like ByteDance, Tencent, Baidu, and Alibaba swiftly followed with price reductions, even cutting prices to below cost margins. This fierce competition stems from minimal technical differentiation between models and slower-than-expected productization. 

From the launch of ChatGPT to July 2024, 78,612 AI companies have either been dissolved or suspended. The competition is not only pushing out the players from the ring, survivors are also drilling down to the niche to differentiate from the others. For example, the industry-specific LLMs are gaining traction, with a significant push from the government. The March 5, 2024 “Government Work Report” delivered by the Chinese Premier minister emphasized the "AI+" strategy,  driving AI’s penetration across industries. By July 2024, the number of AI models registered with the Cyberspace Administration of China (CAC)  exceeded 197, nearly 70% were industry-specific LLMs, particularly in sectors like finance, healthcare, and education. The peace will not last long, AI's rapid integration into vertical industries is expected to become a key area of another round of competition in the coming months.

Under this circumstance, going abroad seems to be a way out.

Pressures from Policy and Investment Environment 

Government is not only incentivizing, but also regulating. Between March and September 2024, the government introduced a series of regulatory policies, particularly around data privacy, algorithm transparency, and content labeling.

• [March 5, 2024]: The China National Information Security Standardization Technical Committee (TC260) released a technical document outlining basic safety requirements for generative AI services. The document mandates security assessments before collecting training data for generative AI models.

• [September 14, 2024]: The Cyberspace Administration of China (CAC) proposed new regulations for labeling AI-generated content. This policy is aimed at ensuring transparency in digital content creation, allowing users to easily distinguish between human-generated and machine-generated material. The primary focus is on mitigating the risks associated with deep fakes, misinformation, and the potential misuse of generative AI technologies.

Regulations are indispensable for any new industry, however they also increase compliance costs for companies, especially for small and medium-sized enterprises. Former Microsoft engineer Shao Meng commented, "Tighter regulations, especially for to-C teams, may push more companies to expand overseas, including their products and even their teams."

On top of the policy pressure, the investment environment is getting more and more rational over the last 6 months compared to the AI fever when ChatGPT was out. 

By mid-2024, Chinese AI startups raised approximately $4.4 billion across 372 funding rounds, a significant drop from the peak in 2021, when investments reached $24.9 billion. 

Overseas Markets, promising land for Chinese AI companies

Compared to the domestic market, one particular element in certain overseas markets is that the individual customers have a greater willingness to pay, thanks to the healthy business environment. By proposing groundbreaking AI solutions meeting the local needs, Chinese AI companies can quickly develop stable revenue streams. For instance, in Southeast Asia, innovative approaches like AI-powered digital human livestreaming are breaking into the e-commerce live-streaming sector.

As for enterprise or government clients, emerging markets like Southeast Asia, the Middle East, and Africa have become the primary choices for Chinese AI companies as mentioned above. These regions, still in the early stages of digital transformation, are jumping directly to the latest technologies . Compared to saturated Western markets, these areas have less competition, higher potential for growth, and lower entry barriers, where Chinese AI tech giants are expanding their market share by capitalizing on their technological strengths, cost-efficient structures, and government support.

What are the key success factors ?

Localization

Regulatory Localization: China has relatively strict AI governance policies, however it focuses more on content safety. While going abroad, Chinese AI companies must navigate diverse data privacy, security, and ethical regulations worldwide, which comes even before the implementation of their business model. EU’s AI Act and privacy protection laws, is a perfect example for Chinese companies to adjust their AI models to meet the EU’s privacy-by-design principles, where data protection is built into the core of AI products and services.

Technical Localization: Despite the magic of AI, there is still no one size fits all solution. In emerging markets with weaker infrastructure, companies need to adjust their products to accommodate network conditions, data storage, and algorithm adaptability. Meanwhile, in developed markets, complying with industry standards such as data localization and ISO certifications etc. 

Boosting International  Influence

Despite the fast growing AI innovation in China, Chinese AI companies have not yet gained enough awareness in overseas markets. Releasing open-source projects became a quick win to build awareness. For example, at least one model from China appears on Hugging Face’s trending model leaderboard almost every one to two weeks. These include Alibaba’s Qwen series, which has been a “long-running hit” on Hugging Face’s Open LLM leaderboard, considered today to be one of the best open LLM in the world; DeepSeek’s coder is another one, that is highly praise by the open source community; and Zhipu AI’s (the profile) also open sourced its GLM series and CogVideo. 

Through open-source initiatives, these projects have gained considerable influence in the international open-source community, helping to enhance recognition, trust, and competitiveness for the Chinese projects in the global market.

An interesting point is that many Chinese companies, after expanding overseas, tend to adopt a new brand name or prefer to promote themselves using the name of their models or applications. “Chinese companies often create new brands for oversea products, even one per country, while Western companies prefer to use unified product names globally,” said Tiezhen Wang, an engineer from Hugging Face. This approach helps them fit into local markets better and shields them from geopolitical pressure at the same time. 

Promoting Environmental, social, and governance (ESG) Strategy

AI for Good (China's proposal for global AI governance) is no doubt an important initiative to explore the potential of AI for a bigger purpose, which is an all inclusive statement without borders. In Beijing, the China ESG30 Forum released the "2024 China Enterprises Global Expansion Strategy Report." This report highlighted the importance of ESG and AI, as two pillars for Chinese companies to integrate into a new phase of globalization. Some tech giants have already begun adopting green energy to drive the sustainable development of their global data centers, or using AI image recognition technologies to monitor wildlife, among others. AI applications are also being used with AI startups and traditional industries to co-develop green technologies, such as renewable energy and electric vehicles. Such innovations further promote product sustainability, helping Chinese firms stand out in the competitive landscape.

Conclusion

Chinese AI companies are at a critical turning point. Expanding overseas is not just a simple market expansion strategy but a necessary choice, for a harsh domestic environment but also a seems-to-be promising overseas opportunities. However, overseas expansion is not guaranteed for success. Under unfamiliar markets and audiences, to be able to quickly adjust to the local market, comply with regulations and build awareness seems also no less challenging.

What’s more, AI is still in an early stage of development, and its true power is unleashed when AI companies find the sweet spot of being an AI+ enabler to reshape the industries. Going abroad is relevant today for Chinese AI companies to grow, but it would become even more relevant when it actually integrates and brings value to the local industries.

Zheng He’s expedition to the “west ocean” was powered by a whole nation strategy thanks to its strong economic power. History seems to repeating itself today but with a different context: Technological innovation thrives not through centralized national efforts, but through the dynamic forces of the free market, where competition, entrepreneurship, and open exchange drive creativity and progress. China’s AI companies have made a long way to rise, and it still needs a long way to flourish.

Thanks to Tiezhen Wang, Luke Cheng, Shao Meng and Sam Guo for providing valuable feedback.

Bonus: Highlighting Key Events (past 6 months of 2024)

• March 5: During the second session of the 14th National People's Congress, Premier Li Qiang delivered the Government Work Report on behalf of the State Council, proposing to deepen research and application in areas such as big data and artificial intelligence, and to launch the "AI+" initiative.

• March 18: Scholars release draft "Artificial Intelligence Law (Scholar's Suggestion Draft)".

• April 3: China opens consultation into data security specifications for generative AI training.

• April 25: The 2024 Zhongguancun Forum (ZGC Forum) held in Beijing, focusing on cutting-edge fields such as AI, life sciences, and new materials. A series of major scientific and technological achievements unveiled during this period.

• May 7:  During the state visit of the President of the People's Republic of China to France, a Joint Statement on Artificial Intelligence and Global Governance was issued by the People's Republic of China and the French Republic.

• May 10: The price war for LLMs has begun.

• May 14: The first meeting of the China-U.S. Governmental Dialogue on Artificial Intelligence was held in Geneva, Switzerland.

• July 4: The World Artificial Intelligence Conference was held in Shanghai, where the Shanghai “Declaration on Global AI Governance” was released.

• August 21: World Robotic Conference was held in Beijing.

• September 9: AI Safety Governance Framework version 1.0 released.

• September 14: CAC proposes new regulation on labeling AI-generated content.

• September 19: Apsara Conference, an annual technology event hosted by Alibaba Cloud, a subsidiary of Alibaba Group. The conference focuses on the latest advancements in cloud computing, artificial intelligence (AI), big data, and other emerging technologies.

Thank you for reading!

Share with at least three of your peers and receive one month of Premium subscription for free 🤍