Zhipu AI Co-Founder Tang Jie: Tokens Are the Engine of China’s Intelligent Economy

Zhipu AI Co-Founder Tang Jie Frames Tokens as the Core Engine of China’s Intelligent Economy

Qiushi, the theoretical journal of the Communist Party of China Central Committee, has published a policy-oriented article by Tang Jie, co-founder of Zhipu AI, arguing that artificial intelligence tokens should be understood as the engine of high-quality development in the intelligent economy.

Driving High-Quality Development of the Intelligent Economy with Tokens as the Engine

The article is not a product announcement. It places a technical measure — the token — at the center of a broader economic and industrial policy discussion, signaling how Chinese AI leaders are trying to define the next phase of AI-driven growth.

What tokens mean in this context

In modern AI systems, tokens are the discrete units of text, code or other data that models process when generating or analyzing content. They are commonly used to quantify model usage, calculate computational cost and manage billing for AI services.

Tang Jie’s article appears to use the term both technically and strategically. Rather than treating tokens only as a measure of model activity, the piece frames token throughput, cost and efficiency as a foundation for productivity gains across industries. That framing matters because tokens are where AI capability meets real-world computing demand: every generated word, line of code or analytical result consumes tokens, and therefore compute, energy and infrastructure.

Notably, this use of tokens is distinct from cryptocurrency. In AI policy discussions, tokens are not financial assets; they are processing units. The distinction is important because the article is focused on productivity and industrial transformation rather than digital currency or blockchain.

A policy narrative, not a launch

The publication venue is notable. Qiushi typically carries theoretical and policy discussions rather than corporate marketing or product news. The choice of venue suggests the article is intended to influence or reflect high-level thinking about AI deployment, industrial coordination and economic planning in China.

The intelligent economy is a term used in Chinese policy discourse to describe AI-enabled productivity, digital transformation and industrial upgrading. It often encompasses:

• Computing power and data infrastructure
• Large-model development and deployment
• Integration of AI into manufacturing, services and public administration
• Measurement and improvement of AI efficiency

If tokens are treated as a core economic unit, they could become a lens for evaluating how AI value is created, distributed and scaled across sectors.

Why the framing may matter for policy

The article could be read as an early signal of alignment between Chinese technology companies and policy bodies around a shared vocabulary for AI’s economic role. It may also point toward future guidance on compute allocation, model deployment standards or industrial AI benchmarks tied to token efficiency.

The discussion arrives as Chinese AI companies face pressure to improve efficiency under hardware constraints. Advanced chips remain a bottleneck for many model developers, making token efficiency — how much useful output can be produced for a given amount of compute — a strategic priority.

China has been pushing to strengthen domestic computing capacity, data circulation and large-model adoption. In that context, positioning tokens as an engine of the intelligent economy could support arguments for treating AI compute and model-serving infrastructure as strategic assets rather than ordinary commercial inputs.

Zhipu AI is one of China’s prominent AI developers, and the presence of a co-founder in Qiushi reflects a broader pattern in which Chinese technology executives contribute to policy-oriented discussions on emerging technology.

The road ahead

The full text of Tang Jie’s article was not immediately detailed beyond its title and stated framing. But the emphasis on tokens as an economic engine suggests the conversation in China is moving beyond raw model capability toward the infrastructure, measurement and policy architecture needed to support large-scale AI adoption.

In industrial terms, the intelligent economy could include AI-assisted manufacturing, automated logistics, smart city systems, financial risk modeling and public service delivery. Each of these applications depends on reliable, affordable token processing at scale.

If the token-as-engine concept gains traction, it could shape how policymakers, companies and regulators think about AI costs, efficiency and industrial value — and how China measures progress in the intelligent economy.