The Rise of China’s ‘AI Tigers’

For years, the United States has held a commanding lead in artificial intelligence research and development. However, recent months have witnessed a dramatic shift, as Chinese firms unveil increasingly sophisticated AI models at an astonishing rate. This surge isn’t merely incremental; it’s a potential inflection point in the global AI landscape.

The momentum began last year with DeepSeek’s R1 model, a surprise release that offered performance comparable to OpenAI’s ChatGPT at a significantly lower cost. This immediately raised questions about the efficacy of US export controls aimed at limiting China’s access to advanced computing power. Now, that initial spark has ignited a full-fledged race.

Moonshot AI and Alibaba Lead the Charge

Just this week, Beijing-based startup Moonshot AI announced Kimi K2.5, a multimodal AI capable of simultaneously processing text, images, and video. But the true innovation lies in its “agent swarm” feature, allowing users to deploy up to 100 independent AI agents to tackle complex tasks in parallel – a capability that promises to revolutionize workflow automation.

Hot on Moonshot’s heels, e-commerce behemoth Alibaba launched Qwen3-Max-Thinking. The company asserts this model isn’t just competitive, it’s superior. According to Alibaba, Qwen3-Max-Thinking outperformed leading US models on “Humanity’s Last Exam,” a notoriously challenging benchmark designed to assess advanced reasoning and problem-solving skills.

“It seems that, for these Chinese start-ups, it was just a matter of getting access to capital,” explains Kyle Chan, a fellow at the Brookings Institution, in a report by SCMP. “I wouldn’t have guessed that Chinese AI companies would continue to keep pace with their US peers as recently as a month or two ago.”

Beyond the Headlines: Baidu, Z.ai, and the Expanding Ecosystem

The acceleration isn’t limited to Moonshot and Alibaba. Z.ai recently released a free version of its GLM 4.7 model, though demand quickly overwhelmed its computing resources, forcing a temporary halt to new sign-ups. Baidu, meanwhile, has seen its Hong Kong-listed shares reach a nearly three-year high following the unveiling of its latest Ernie 5.0 model.

The Power of Open Source

A key differentiator between the Chinese and American approaches to AI development is the embrace of open-source models. While companies like OpenAI and Anthropic largely guard their underlying technology, Chinese firms are actively releasing the “weights” of their models, allowing developers worldwide to freely customize and build upon their work.

This strategy has significant cost implications. Artificial Analysis, as cited by SCMP, estimates that Moonshot’s K2.5 is more than four times cheaper to operate than comparable models from OpenAI. This affordability is already attracting attention from Western tech giants, with Airbnb recently adopting Alibaba’s Qwen for its customer service AI and Pinterest integrating Chinese models into its recommendation engine.

Did You Know? The open-source approach allows for faster iteration and broader collaboration, potentially accelerating AI development globally.

Circumventing the ‘Silicon Noose’

Despite US restrictions on the export of advanced Nvidia chips – essential for training large AI models – Chinese innovation has not been stifled. Instead, engineers are employing creative solutions, such as maximizing “Mixture-of-Experts” architectures, to achieve more with limited computing power.

Furthermore, China is investing in domestic chip production. Zhipu AI recently launched GLM-Image, claiming it’s the first model fully trained on Chinese-made chips. Even Google DeepMind CEO Demis Hassabis acknowledges the shrinking gap, telling CNBC that China’s AI capabilities may be “months” behind those of the US.

AI Beyond the Digital Realm

The competition extends beyond software. Tencent is integrating AI into everyday life, offering $140 million in cash rewards through its Yuanbao AI app during the Lunar New Year. XPeng founder He Xiaopeng recently showcased the humanoid robot IRON, demonstrating such realistic movement that he had to publicly disassemble its leg to prove it wasn’t a human in disguise.

As these “AI Tigers” continue to attract billions in investment – with Moonshot recently achieving a valuation of nearly $4.8 billion – the focus is shifting from simply building AI to rapidly integrating it into the global economy. What impact will this have on industries worldwide, and how will governments respond to this evolving landscape?

Will the open-source model prove to be a more effective path to innovation than the closed-garden approach favored by some US companies? And what are the long-term implications of a two-tiered AI ecosystem, with potentially different standards and values?

Frequently Asked Questions About China’s AI Advancement