The “AI Tigers” Boom: China’s LLM Revolution
In the fast-evolving world of artificial intelligence, China’s startup ecosystem is entering a new era — one defined by homegrown large language model (LLM) pioneers often dubbed the “AI Tigers.” Among this elite group, three companies Zhipu AI, MiniMax, and Moonshot AI — have emerged as the most visible contenders in a competitive race to define China’s future in foundational AI. Together, they illustrate both the technological ambition and market pressure shaping the next phase of global AI innovation.
What Are the “AI Tigers”?
The term “AI Tigers” refers to a cluster of leading Chinese AI startups focused on developing large language models and generative AI technologies. These firms stand alongside other major players such as Baichuan Intelligence, 01.AI, and StepFun, frequently making headlines for their technological advances, strategic fundraising, partnerships with global tech giants, and increasingly their efforts to go public.
This cohort is seen as a counterweight to both Western AI leadership (e.g., OpenAI, Google) and newer domestic rivals like DeepSeek, whose open-source models have disrupted the landscape with low-cost, high-performance alternatives.
Zhipu AI: From Startup to IPO Contender
Zhipu AI (also referred to as Knowledge Atlas Technology) is one of the earliest and most ambitious Chinese AI model builders, founded in 2019 out of academic roots tied to Tsinghua University’s Knowledge Engineering Group.
Growth and Strategy
Zhipu has raised significant capital, attracting backing from major tech firms like Alibaba, Tencent, Xiaomi, and Meituan, as well as sizable state-linked investment funds. In 2025, despite being added to the U.S. export control list, the company secured over a billion yuan in state and private funding, underlining Beijing’s strategic support for domestic AI capacity.
Zhipu’s core focus remains on building a foundation model (GLM) architecture with ambitions not only to match global standards but also to serve enterprise clients at scale. The company has indicated intentions to open-source parts of its model suite, including foundation models, inference systems, multimodal models, and AI agents to sustain competitiveness.
The IPO Moment
By late 2025, Zhipu had filed for an IPO on the Hong Kong Stock Exchange, positioning itself as possibly the world’s first public LLM company ahead of U.S. pioneers like OpenAI and Anthropic. Its stock debut in January 2026 attracted global attention, though its price increase was more modest compared with peers.
MiniMax: Consumer, Commerce, and the Public Market
MiniMax, founded in 2021 and headquartered in Shanghai, stands out for its strong consumer-facing applications built atop LLM research.
Its early product efforts include AI chat platforms such as Talkie (Xingye), which allows conversational engagement with various characters and generative personas a strategy that has helped drive user growth and brand recognition domestically and internationally.
Strategic Backing
MiniMax benefits from heavyweight investors such as Alibaba and Tencent, giving it an edge in both ecosystem integration and marketplace traction. However, it has also faced growing pains, including an intellectual property lawsuit from major Hollywood studios in the U.S., highlighting broader IP tensions as AI scales.
IPO Breakout
MiniMax made a spectacular debut on the Hong Kong stock market in early 2026, with its shares surging more than 100% on day one. The company raised roughly HK$4.8 billion (~$620 million), signaling strong investor confidence in China’s generative AI story particularly among firms marrying LLM technology with user-centric apps.
Moonshot AI: Intelligence at Scale
Moonshot AI offers a slightly different angle in the AI Tigers landscape. Founded in 2023 by Tsinghua and Carnegie Mellon alumnus Yang Zhilin, the company gained rapid visibility with its Kimi AI chatbot — notable for being among China’s top chatbots and for processing very long prompt inputs compared with many rivals.
Funding and Technical Ambitions
Moonshot has raised hundreds of millions in financing, including a major $500 million Series C round, and boasts rapidly growing global paid user metrics. That capital is earmarked for GPU capacity expansion, advanced model R&D (including K3 models), and pushing into more intelligent agents and enterprise use-cases.
What This Means for China’s AI Ecosystem
These three startups reflect different strategic visions within China’s LLM boom:
- Zhipu AI emphasizes foundational model leadership and deep enterprise integration.
- MiniMax blends LLM technology with mass consumer usage and strong capital market performance.
- Moonshot AI aims to scale both technical innovation and global reach via advanced reasoning models and agent frameworks.
Their trajectories also underscore broader trends: China is not just replicating Western AI models, but building distinct commercial and regulatory paths, sometimes bolstered by state support and sometimes propelled by capital markets like the Hong Kong stock exchange.
The race is far from settled with newcomers such as DeepSeek disrupting cost dynamics but the emergence of IPO-ready LLM companies signals a maturing AI industry ready to compete globally.
Final Takeaway
The “AI Tigers” boom led by Zhipu AI, MiniMax, and Moonshot AI marks a pivotal shift in China’s approach to large language models. These companies combine deep technology development, massive funding, tech-giant backing, and market ambition in ways that could reshape not just China’s technology landscape but the global AI terrain as well. As they move from R&D labs to the public market and commercial deployment, the world will watch closely and the second wave of AI competition is just beginning.