From " Selling Products" to "Taking Root": The New Logic of Chinese AI Companies Going Global
Hello everyone, I'm your financial observer. Today, we're going to discuss a profound transformation that's taking place among China's tech giants and unicorns.
If you follow tech news, you've probably noticed a term that's been appearing frequently lately: "going global". In the past, when companies thought about going global, they meant things like "selling phones in Africa" or "exporting electric vehicles to Europe"—the focus was always on selling products.
But things have changed. At a high-level dialogue in Malaysia, several tech leaders and diplomats reached a consensus: Going global now is not just about simple exportation; it's about establishing capabilities and jointly creating value.
In simple terms, it used to be a matter of "I produce, you pay"; now it's "I bring the technology, you provide the context, and together we make the business bigger, with the benefits staying locally."
Let me break down this dialogue into five key points to explain the behind-the-scenes strategies in plain language.
---
1. The Core Shift: From Selling Tools to Teaching How to Use Them
In the past, Chinese companies going global was like selling tools. They would produce phones or servers, ship them abroad, and once the money was received, the transaction was complete. But with the advent of AI, the approach has changed. Li Zhu, from Facewall Intelligence, pointed out that modern AI, especially large-scale models, doesn't need to process all data in distant cloud centers; it can perform computations directly on your phone, car, robot, or even within your company's internal network.
What does this mean?
It means AI can become much more compact and can be directly integrated into local devices. For countries like Malaysia, where infrastructure is still being developed rapidly, this represents a huge opportunity. There's no need to spend billions on building supercomputing centers; existing hardware can be used to upgrade systems intelligently.
So, the role of Chinese companies has shifted from being mere product suppliers to becoming providers of capabilities. We bring the AI "engine" to the local environment, enabling local devices and networks to function more efficiently. It's like we used to sell flour; now we teach local bakeries how to mix and ferment the dough, and even help them improve their recipes to make better bread.
In one sentence: The focus of going global has shifted from selling products to integrating capabilities.
---
2. The Survival Rule: Don't Compete Head-On with Giants; Focus on Local Integration
Many startups making their global debut make the mistake of trying to directly compete with local giants. This often ends in failure. Su Liangliang from Yufan Intelligence proposed a practical strategy: saturation attack and deep integration. Startups should focus on a specific niche and excel in it to build a competitive advantage. More importantly, avoid direct competition with end-users. Local customers tend to trust local service providers.
Su Liangliang's approach is to package his algorithms and product platforms with local integrators (such as IT or engineering firms). These partners understand the local language, culture, and regulations better and are closer to the customers. They define the use cases and provide services, while he provides the underlying technology.
To put it another way: Instead of opening a restaurant and selling food directly, I license my unique sauces and cooking techniques to dozens of local restaurants. The customers enjoy the food, but I earn from the licensing and long-term cooperation.
The key point: The technology can be imported, but the capabilities must be localized. This way, the business becomes a joint effort between the two countries, not just one company's success.
---
3. The Value Loop: The Model Isn't the Endgame; Practical Use Matters
There's a misconception in the AI industry that larger model sizes and higher scores equate to better performance. Su Liangliang and Li Zhu both emphasize that the model itself is no longer the deciding factor. The real competition is about transforming model capabilities into products that customers can perceive, use, and benefit from.
For example, RytBank, a digital bank in Malaysia, gained 1.5 million users in just 7 months because it used Ant Group's eKYC (electronic identity verification) and anti-money laundering solutions. The key here is that Chinese technology solved the problem of quickly and securely identifying users, while the local ecosystem (YTL AI Labs) ensured that locals were willing and able to use the services.
In other words: Having a smart model is important, but it needs practical applications and local support to be truly valuable.
---
4. Historical Lessons: From Semiconductors to AI, Malaysia Is the Perfect Launchpad
Why Malaysia, and why ASEAN? The former Malaysian Ambassador to China, Datuk Muhid, shared a meaningful story: In the 1970s, when multinational semiconductor companies entered Malaysia, they didn't just see it as a market; they brought technology and trained local talent. As a result, Malaysia became a major semiconductor manufacturing hub and developed a skilled workforce capable of collaborating on product development.
What does this history teach us about going global with AI?
1. Focus on more than just the market; focus on building a foundation: Malaysia has an AI security network and national-level AI institutions, providing a conducive policy and governance framework, as well as a talent pool.
2. Language and cultural advantages: Malaysia is a bilingual country with a relatively open culture, making it an ideal entry point for Chinese companies entering the entire ASEAN region.
3. Joint development is key: Cooperation should be about mutual creation, not just one-way transactions.
In one sentence: Malaysia is not just a starting point; it's a strategic base for expanding into the entire ASEAN.
---
5. Ecological Collaboration: Policy, Capital, and Talent Are All Essential
Going global requires a comprehensive ecosystem. Datuk Muhid and Yang Shoujun outlined the future vision from policy and capital perspectives:
Policy Measures:
- Four Corridors:
- Technology Corridor: Attracting Chinese AI companies to set up regional headquarters or R&D centers in Malaysia.
- Talent Corridor: Jointly training engineers to address the talent shortage.
- Innovation Corridor: Establishing joint laboratories for local application development.
- Investment Corridor: Creating Sino-Malaysian funds to support local startups.
Capital Requirements:
Different stages of the AI industry require different types of funding:
- Basic Research: Long-term capital is needed because returns are not immediate, and government or patient capital is essential.
- Technology Transformation: Patient capital is required to overcome the challenges of the technology-to-product transition.
- Commercialization: Industrial capital is needed to support companies during the growth phase.
The ultimate goal: The goal is to use the success of unicorn companies to drive the development of the AI industry in parks, cities, and throughout ASEAN.
---
Conclusion: The Next Phase is a Race of Long-Termism
This dialogue clearly signals the arrival of a new era for Chinese tech going global:
- Phase 1.0: Focused on products, prices, and channels, with the main goal of selling.
- Phase 2.0: Emphasizes building ecosystems, localization, and joint creation, with the focus on integration.
Success will no longer depend on the advancement of individual technologies but on who can establish a strong presence locally and create value together.
For Chinese tech companies, this represents not just business opportunities but a strategic leap. It's about moving from simply entering the market to truly becoming part of the local ecosystem.
For consumers, this means that AI applications in Southeast Asia will increasingly reflect joint Chinese-Malaysian development and localized services tailored to the region. This is the real opportunity in the AI revolution.