The global race for artificial intelligence dominance has taken a new turn, with OpenAI publicly acknowledging the significant and rapidly accelerating progress of Zhipu AI, a Chinese startup. This isn’t just about technological advancement; it represents a deepening rivalry between the United States and China, and the potential reshaping of the digital world order. Let’s delve into the story of Zhipu, its ambitions, and what its rise means for the future of AI.

Understanding Zhipu AI: A State-Backed Challenger

Recent observations by OpenAI analysts have highlighted Zhipu AI’s remarkable success in securing government contracts across diverse regions, including Malaysia, Singapore, the UAE, Saudi Arabia, and Kenya. This isn’t a coincidence; it’s a deliberate strategy. Zhipu’s mission is clear: to establish Chinese AI systems and standards within emerging markets before their Western counterparts can gain a foothold. This initiative cleverly leverages the Belt and Road Initiative, transforming it into a “Digital Silk Road” strategy.

The Foundation: Government Support and Strategic Partnerships

Zhipu operates in close alignment with the Chinese government and state-owned enterprises. The company has received substantial financial backing, exceeding $1.4 billion in state funding. This support is a key differentiator, allowing Zhipu to offer competitive solutions and pursue ambitious global projects. Furthermore, collaborations with tech giants like Huawei ensure the delivery of sovereign large language model (LLM) infrastructure and private hardware to partner nations. Zhipu positions itself as a “responsible, transparent, and audit-ready” alternative to Western AI, appealing to governments seeking control and oversight.

The Race for Sovereign AI: A Global Competition

The current landscape sees both Zhipu and OpenAI targeting the same prize: countries striving to build their own “sovereign AI” capabilities. Sovereign AI signifies locally controlled and customized AI systems tailored to national priorities. Zhipu’s strategy centers on locking in Chinese systems before U.S. or European tech can establish a significant presence. This mirrors OpenAI’s own “OpenAI For Countries” initiative, demonstrating a direct and escalating competition. It’s a pivotal battle in a broader contest for influence over the next wave of digital infrastructure.

Technological Innovation, Open Source Strategies, and Global Aspirations

Founded in 2019, Zhipu emerged as one of the first Chinese firms to develop independent large language models that move beyond architectures based on GPT. The company’s GLM series showcases both proprietary and open-source models, including GLM-4-9B and GLM-4V-9B, which offer sophisticated text and multimodal capabilities. The business model revolves around Model-as-a-Service (MaaS), allowing clients to fine-tune LLMs for specific needs. The strategic release of some models under open-source licenses is intended to foster developer engagement and innovation.

Multimodal AI and the Sora-Like Generator

The July 2024 launch of a text-to-video generator comparable to Sora underscored Zhipu’s ambition to compete head-to-head with Western leaders in the field of multimodal AI. While its GLM-4-Plus model currently ranks as the highest-performing Chinese-developed LLM, it still lags behind top U.S. models in complex reasoning and adaptability. However, Zhipu’s rapid ascent, from academic research to a major force within China’s AI ecosystem, highlights the power of university-driven innovation within the country.

Fueling Expansion: Strategic Funding and a Global Push

Zhipu’s expansion isn’t solely reliant on state funding. It’s a combination of government support and international investment. A recent $400 million funding round, including participation from a unit of Saudi Aramco, speaks volumes about international confidence in Zhipu’s potential. Despite being added to the U.S. export control list in January 2025, the company’s momentum hasn’t slowed. Instead, Zhipu is aggressively pursuing partnerships across Asia, Africa, and the Middle East, aiming to establish Chinese AI as the default solution in emerging markets.

The Intensifying U.S.-China AI Rivalry

OpenAI’s recent caution regarding Zhipu’s global ambitions is occurring during a period of increasingly fierce competition between the U.S. and China for AI leadership. Both sides are actively vying for government contracts and working to shape international standards. The outcome of this contest will significantly influence not only the future of AI technology but also the geopolitical balance of power in the digital age.

Timeline of Zhipu’s Key Milestones:

• 2019: Zhipu AI is founded.
• 2024 (July): Launch of a text-to-video generator comparable to Sora.
• 2025 (January): Added to the U.S. export control list.
• [Ongoing]: Aggressive pursuit of partnerships in Asia, Africa, and the Middle East.

The Takeaway: A New Chapter in the Tech Rivalry

Zhipu AI’s aggressive global expansion, sophisticated sovereign AI infrastructure, and well-planned open-sourcing strategy have positioned it as a crucial player in China’s pursuit of AI leadership. As it gains traction in emerging markets, the world is witnessing a new chapter in the U.S.-China tech rivalry—a battle for AI supremacy fought not only within laboratories, but also in the halls of government and the fundamental infrastructure of nations. The implications are far-reaching and will shape the technological and geopolitical landscape for years to come.