China's Big 3 Launch Token-Based AI Services

China’s three major telecommunications operators have officially launched token-based AI services on the National Computing Power Platform. This move marks a significant shift toward standardized pricing and billing for large language model (LLM) usage in China.

The China Academy of Information and Communications Technology (CAICT) announced on June 3 that China Telecom, China Mobile, and China Unicom have integrated their 'Token Products' into the national infrastructure. This development aims to streamline AI adoption for businesses and individual users alike.

Key Facts: The New AI Billing Standard

• Standardized Unit: The 'Token' is now the official minimum unit for measuring, pricing, and trading AI information processing.
• Major Players: China Telecom, China Mobile, and China Unicom are the first to adopt this unified billing framework.
• Platform Integration: All services are hosted on the China Computing Power Platform, ensuring centralized access and management.
• Commercial Focus: The initiative targets both enterprise developers and individual consumers with tiered service plans.
• Cost Reduction: A primary goal is to significantly lower the barrier to entry for AI services through transparent pricing.
• Model Diversity: Plans include access to GLM-5, DeepSeek V3.2, and various coding-specific models.

Understanding the Token Economy

In the context of artificial intelligence, a token represents the smallest unit of information processed by a large language model. Unlike traditional cloud computing metrics such as CPU hours or storage gigabytes, tokens directly correlate with the complexity and volume of AI interactions.

This new value system treats AI inference as a tradable commodity. By quantifying AI output in tokens, providers can offer precise pricing models. This transparency helps enterprises predict costs more accurately compared to opaque subscription fees or flat-rate API access.

The evolution of this token economy mirrors the early days of electricity metering. Just as utilities moved from flat rates to per-kilowatt-hour billing, AI providers are shifting toward granular usage-based pricing. This allows for fairer compensation for compute resources while enabling users to pay only for what they consume.

China Telecom’s Tiered Approach

China Telecom has introduced its Tianyi Cloud Token Plan, which segments users into two distinct categories. The first category targets developers and small-to-medium enterprises (SMEs). It leverages the capabilities of the GLM-5 large model.

This enterprise tier focuses on professional needs such as code development, complex logical reasoning, and long-text processing. It supports lightweight business development and batch calls, making it ideal for high-frequency commercial scenarios. Users can integrate these tokens into daily project iterations without worrying about massive upfront infrastructure costs.

The second category serves individuals and households. It utilizes the DeepSeek V3.2 general-purpose large model. This plan offers lightweight token products designed for everyday tasks. These include office assistance, creative writing support, and lifestyle consultations. This segmentation ensures that casual users do not overpay for enterprise-grade computational power.

China Mobile’s Coding-Centric Strategy

China Mobile has taken a specialized approach with its Coding Plan. This product is specifically engineered for software developers seeking AI-assisted programming tools. It functions as an AI coding subscription service.

The service integrates mainstream code models and maintains compatibility with popular AI programming environments. This ensures a smooth workflow for engineers who rely on tools like VS Code or JetBrains IDEs. The coverage extends beyond simple code generation to include code review and architectural design.

By focusing on coding, China Mobile addresses a high-value niche within the developer community. Automated code generation reduces debugging time and accelerates deployment cycles. This targeted strategy differentiates them from generalist LLM providers who may lack deep integration with development workflows.

Industry Context and Global Comparison

This initiative places China at the forefront of structured AI commerce. While Western companies like OpenAI and Anthropic charge per token, they operate in a fragmented market with varying standards. China’s state-backed platform creates a unified marketplace.

Compare this to the US market, where AWS, Azure, and Google Cloud compete with independent AI firms. In China, the telecom giants dominate the infrastructure layer. Their collaboration via the CAICT platform suggests a coordinated national strategy to accelerate AI industrialization.

The emphasis on 'tokens' as a currency highlights the commoditization of intelligence. As models become more accessible, the competitive edge shifts from raw model performance to ease of integration and cost efficiency. This trend is likely to influence global pricing strategies as the market matures.

What This Means for Developers and Businesses

For businesses, the immediate benefit is cost predictability. The ability to purchase tokens in bulk or on-demand allows for better budget management. SMEs can now experiment with advanced AI features without committing to expensive custom solutions.

Developers gain access to standardized APIs across multiple providers. This interoperability reduces vendor lock-in risks. If one provider raises prices, users can potentially switch to another on the same platform with minimal friction.

Individual users also benefit from lower entry barriers. The household plans make sophisticated AI assistants affordable for daily use. This democratization of AI tools could lead to widespread adoption in education and personal productivity sectors.

Looking Ahead: Future Implications

The launch of these token products is just the beginning. We can expect further refinement of pricing models based on real-time usage data. Future iterations may introduce dynamic pricing during peak demand periods.

Additionally, the success of this platform could inspire other countries to adopt similar centralized computing exchanges. It sets a precedent for how national infrastructure can support emerging technologies.

Watch for expansion into other AI modalities, such as image and video generation. The token concept may eventually apply to all forms of generative AI, creating a universal metric for digital creativity.

Gogo's Take

• 🔥 Why This Matters: This moves AI from a 'black box' expense to a measurable utility. For Western observers, it signals China’s intent to standardize AI commerce at a national scale, potentially creating a more efficient, albeit controlled, market compared to the fragmented Western landscape.
• ⚠️ Limitations & Risks: Centralization poses risks regarding data privacy and vendor dependency. If the National Computing Power Platform experiences outages or policy changes, millions of users could be affected simultaneously. Additionally, reliance on domestic models may limit access to cutting-edge global innovations.
• 💡 Actionable Advice: Developers should monitor these token pricing structures as benchmarks for global AI costs. If you are building AI-integrated apps, consider diversifying your model providers to avoid lock-in. Keep an eye on how these 'token' definitions evolve, as they may become the industry standard for billing.