Zhipu's Tmall Token Retail Billing Model
1) Token retail: retail packages billed by actual Token consumption through online channels such as Tmall; API revenue s
Key Fields
FIELD STAMPS📌 Background
The share of large model API revenue is rising rapidly, and Token consumption has become a core metric for enterprise payment. Zhipu's semi-annual report shows first-half revenue of RMB 954 million, up 399.7% year over year, with open platform and API revenue of RMB 825 million, whose share rose from 15.2% in the same period last year to 86.5% (as reported in the company's semi-annual report). During the same period, China's average daily Token consumption exceeded 140 trillion, more than 1,000x higher than at the beginning of 2024. Based on this, Zhipu moved its Coding Plan to Tmall, opening a store to sell usage volume.
👤 Target Customers
Enterprise developers, small and medium-sized enterprises, and individual developers using large model APIs
💰 Revenue Streams
1) Token retail: retail packages billed by actual Token consumption through online channels such as Tmall; API revenue share rose from 15.2% to 86.5% (merchant self-reported, not independently verified); 2) Usage tiers: tiered pricing for enterprise customers based on API call volume; 3) Package subscriptions: periodic subscription fees for specialized packages such as coding for developers; 4) Private deployment: localized model deployment for enterprise customers, charged per project (opportunity item; no figures available on how much revenue it could generate after implementation).
🧮 Cost Structure
Compute infrastructure costs, model inference costs, Tmall platform channel fees, sales and operating costs
🛡️ Moat
Model performance and API ecosystem lock-in; high migration costs after a large number of enterprise customers have integrated. Platform-based channels lower customer acquisition costs and create economies of scale.
🔑 Keys to Success
- Token pricing matches costs
- Retail channels such as Tmall lower the purchase barrier
- API ecosystem locks in enterprise customers
⚠️ Risks
- Rising compute costs erode gross margin
- Competitor price cuts or outcome-based pricing disrupt usage-based billing
🏢 Cases
- Zhipu opened an official store on Tmall to sell Token packages
- Zhipu's API revenue share rose from 15.2% in 2025 to 86.5% in 2026
📊 SWOT Analysis
Strengths
- API revenue share exceeds 80%, sufficiently validating commercialization
- Token usage grew 40x, showing strong customer stickiness
Weaknesses
- Dependence on compute costs; gross margin is affected by chip price fluctuations
- Retail-oriented channels may lower average order value
Opportunities
- Enterprise AI budgets are shifting from subscriptions to usage-based billing
- Tmall channel reaches the long-tail developer market
Threats
- Competitors such as OpenAI adopting outcome-based pricing, creating price wars
- Open-source models push down Token unit prices