Zhipu AI’s ARR Hits $1.6B, Targets $2.4B by Year-End; Admits Lagging Anthropic by 17 Months
TREE NEWS reports: Chinese AI startup Zhipu AI (智谱) released its first interim results since its Hong Kong IPO, revealing a rapid acceleration in annual recurring revenue (ARR) and a bold year-end target. During an earnings call on August 31, the company disclosed that ARR reached $1.6 billion at the end of August, with an official year-end guidance of $2.4 billion. This marks a significant jump from the $250 million reported in March and $1 billion in July. The company also candidly acknowledged that it still trails U.S. rival Anthropic by about 17 months in ARR terms, though it has narrowed the gap from 24 months in just seven months.
Key Metrics: Price Hikes and Volume Surge
Despite a 101% average increase in API prices, token call volumes have grown more than 40-fold since the start of the year. The top ten customers by revenue have increased their daily token usage by 98-fold. The company’s Coding Plan subscription product saw a 234-fold increase in usage over 365 days, even after a price hike and a period of restricted sales due to compute shortages. This ‘price-volume surge’ suggests that demand for Zhipu’s models is highly elastic, driven by improved model capabilities rather than price discounts.
Market Impact and Analysis
The news underscores the intensifying competition in the global AI race. Zhipu’s aggressive growth in ARR and its explicit comparison to Anthropic signal that Chinese AI firms are closing the gap with U.S. leaders, despite export controls on advanced chips. For investors, this has several implications:
- AI Sector Sentiment: Zhipu’s strong performance could boost sentiment for AI-related stocks, especially those with exposure to Chinese AI infrastructure and applications. However, it also highlights the high capital expenditure and compute constraints that AI companies face.
- Compute and Chip Supply: Zhipu’s reliance on domestic chips for its GLM-5.3 Flash model (running 62 trillion tokens in six days on domestic hardware) validates the progress of Chinese chipmakers like Huawei and Cambricon. This could positively impact Chinese semiconductor stocks and negatively affect expectations for Nvidia’s dominance in China.
- Cloud and Enterprise Spending: The rapid adoption of Zhipu’s models by top Chinese internet companies (4 of the top 10 have made GLM their primary global choice) signals a shift in enterprise AI spending towards domestic solutions, potentially impacting both Chinese and U.S. cloud providers.
Key Takeaways for Investors
- Zhipu’s ARR growth trajectory is impressive but still lags U.S. peers; watch for continued convergence.
- Price hikes without demand destruction indicate strong product-market fit, a positive sign for AI monetization.
- The push for domestic chip usage is a strategic hedge against U.S. sanctions and could redefine supply chains.
- Gross margin guidance of 50% within 12-18 months is ambitious; monitor progress as a proxy for operational efficiency.
Zhipu’s next-generation model, centered on ‘Fully Self-Training’ (autonomous pre-training and post-training with self-correction), could be a differentiator. Investors should watch for updates on this front, as well as the company’s overseas expansion plans, which are expected to materialize within 1-2 months.
