In less than two years, China has produced four AI startups with a combined valuation exceeding 1 trillion yuan (roughly $140 billion). Their rise has become one of the most closely watched developments in the global AI race, especially as China pushes to build domestic alternatives to U.S.-dominated AI infrastructure.

Known in Chinese tech circles as the "AI Four Dragons," these companies are leading China's competition in large language models (LLMs), AI applications, and AI infrastructure.

The four companies are:

• DeepSeek — valued at roughly $45 billion
• Zhipu AI — valued at approximately 377.5 billion yuan
• MiniMax — valued at approximately 223.5 billion yuan
• Moonshot AI (Kimi) — valued at more than $20 billion

Together, they represent China's strongest group of independent AI startups outside the country's established tech giants such as Alibaba, Tencent, and ByteDance.

But the real story is not just their valuations. It is what these companies reveal about how China's AI industry is evolving—and why investors believe only a few players may ultimately survive.

Why China's AI Boom Matters

For years, Chinese technology investors focused primarily on traditional financial metrics: revenue growth, profitability, and market share.

The AI sector is changing that.

Today, many investors are valuing Chinese AI companies based not only on current business performance, but also on their potential to become foundational platforms for the future AI economy.

In Chinese venture capital circles, this is sometimes described as competition for "definition rights"—the ability to shape technical standards, developer ecosystems, and future infrastructure.

The concept resembles how Silicon Valley has valued companies such as OpenAI and Anthropic: investors are betting less on near-term profits and more on long-term platform dominance.

At the same time, AI has become increasingly tied to national strategy in China.

A major signal came when China's National Integrated Circuit Industry Investment Fund—commonly known as the "Big Fund"—reportedly participated in DeepSeek's latest funding round. The state-backed fund historically focused on semiconductor manufacturing and chip supply chains, not AI applications.

Its involvement suggests that advanced AI models are now being treated as strategic infrastructure alongside semiconductors.

Meet China's AI Four Dragons

DeepSeek: China's Strategic AI Challenger

Among the four companies, DeepSeek has attracted the most global attention.

The company built its reputation through open-source models such as V3, R1, and V4, which gained traction within the international AI developer community. DeepSeek also became a symbol of China's effort to reduce dependence on U.S. technology after demonstrating that its models could run effectively on domestic hardware, including Huawei's Ascend 910B chips.

That capability matters because U.S. export controls have restricted China's access to advanced NVIDIA AI chips.

For many investors, DeepSeek's value lies not only in commercial potential, but also in its strategic importance to China's domestic AI ecosystem.

Zhipu AI: China's Enterprise AI Leader

Zhipu AI is often viewed as the most commercially mature company among the Four Dragons.

The company reportedly generated 724 million yuan in revenue in 2025, with more than 240,000 paying developers and partnerships involving 9 of China's top 10 internet companies.

Unlike some rivals that focus heavily on consumer applications, Zhipu has concentrated on enterprise AI services and developer platforms.

Because of this combination of revenue growth and infrastructure ambitions, some investors compare Zhipu to Anthropic in the U.S. market.

MiniMax: Betting on Consumer AI

MiniMax represents China's push into consumer-facing AI products.

The company is known for applications such as Talkie and Xingye, which focus on AI companionship and interactive entertainment. Approximately 70% of its revenue reportedly comes from overseas markets, giving it one of the strongest international profiles among Chinese AI startups.

MiniMax also improved gross margins significantly in 2025, rising from 12.2% to 25.4%.

Investors are betting that consumer AI products could eventually scale to hundreds of millions of users globally, though this remains one of the riskiest segments of the AI market.

Moonshot AI (Kimi): Speed and Product Momentum

Moonshot AI, best known for its Kimi chatbot, has become one of China's fastest-growing AI startups.

The company built its reputation around long-context AI models and reasoning capabilities. After releasing its K2.5 model, Moonshot reportedly doubled ARR month-over-month, surpassing $100 million in March 2026 and $200 million in April 2026.

Its rapid growth has helped position Kimi as one of China's best-known consumer AI brands.

Why Only a Few Companies May Survive

China's AI industry is already consolidating rapidly.

Just a year ago, investors frequently referred to a broader group of AI startups known as the "Six Tigers." Several companies have since retreated from the race to build frontier general-purpose models.

For example:

• Baichuan Intelligence shifted toward medical AI
• 01.AI pivoted toward enterprise services and B2B applications

The reason is simple: building frontier AI models is extremely expensive.

Training costs, inference costs, and access to advanced chips have created enormous financial pressure across the industry.

The Core Battle: Compute, Talent, and Distribution

At its core, China's AI race is increasingly becoming a competition over three things:

• Access to computing power
• Ability to retain top researchers
• Ability to attract users and developers

Each company is approaching the problem differently.

• Zhipu AI focuses on premium enterprise customers
• MiniMax prioritizes rapid user growth and developer ecosystems
• DeepSeek emphasizes architectural efficiency and domestic-chip compatibility
• Moonshot AI relies heavily on software optimization to compensate for hardware constraints

But all four face the same challenge: AI capabilities are converging quickly, while infrastructure costs remain extremely high.

The Commoditization Risk

Another major challenge is product homogenization.

As AI assistants, coding tools, and AI agents become more common, companies are finding it increasingly difficult to maintain differentiation.

After OpenClaw popularized computer-use AI agents in early 2026, several Chinese competitors released similar products within weeks, including Kimi Claw, MaxClaw, and AutoClaw.

The speed of imitation reflects a familiar pattern from China's earlier internet economy: intense competition, rapid replication, and aggressive pricing pressure.

That dynamic can accelerate adoption—but it can also compress margins and weaken long-term profitability.

What Determines the Winners

Several factors are likely to shape the next phase of China's AI industry.

1. Talent Retention

Competition for elite AI researchers has intensified dramatically.

Major Chinese tech companies are offering compensation packages worth tens of millions of yuan, making it difficult for startups to retain top talent without continuously raising capital.

2. Revenue Growth

Strong ARR growth remains essential.

For comparison:

• Zhipu AI reportedly reached roughly 2.5 billion yuan ARR by March 2026
• Moonshot AI surpassed $200 million ARR in April 2026

The key question is whether revenue can grow quickly enough to offset rising compute costs.

3. Domestic AI Chips

China's AI industry is also racing to build a viable domestic hardware ecosystem.

DeepSeek's valuation partly reflects its ability to optimize models for Chinese chips such as Huawei Ascend. However, Alibaba, Huawei, and other large players are building competing AI infrastructure stacks as well.

4. Balancing State Support and Commercial Expansion

Government support provides funding and political backing, but it may also create tensions around commercialization and global expansion.

Chinese AI companies must balance domestic strategic priorities with the need to compete internationally.

What Happens Next?

The future of China's AI sector could unfold in several ways.

In a bearish scenario, AI models become commoditized, similar to cloud infrastructure services. If capabilities converge and price competition intensifies, valuations could compress sharply.

In a bullish scenario, one or two companies successfully establish platform dominance through superior technology, stronger ecosystems, or full-stack integration across chips, models, and applications.

Most likely, the market will continue to diverge.

Each of the Four Dragons is pursuing a different strategy:

• DeepSeek is increasingly viewed as a strategic infrastructure player
• Zhipu AI is building an enterprise AI platform
• MiniMax is betting on consumer AI adoption
• Moonshot AI is racing to convert product momentum into sustainable scale

The trillion-yuan valuation milestone is therefore less a final destination than a snapshot of a rapidly evolving industry.

China's AI race is no longer simply about catching up to Silicon Valley. It is becoming a contest over who will define the next generation of AI infrastructure inside the world's second-largest economy.