Zhipu AI Limits Coding Agent Access After GLM-4.7 Surge

In a remarkable demonstration of the growing demand for AI-powered development tools, Chinese AI firm Zhipu has announced it will limit access to its coding assistant following overwhelming response to its latest GLM-4.7 large language model.

What Happened?

Starting Friday (January 23), Zhipu will accept only 20% of its current daily new subscriptions to its GLM Coding Plan. The company clarified that existing automatic renewals will not be affected by this temporary measure.

The recent launch of GLM-4.7 triggered a surge of users, resulting in temporary computing resource constraints. — Zhipu (WeChat Statement)

About GLM-4.7 and Zhipu

Zhipu's GLM-4.7 is a large language model designed specifically for coding assistance, similar to Anthropic's Claude. The Beijing-based company has positioned itself as a strong contender in the AI space, competing directly with major players like:

• Anthropic (Claude)
• OpenAI (ChatGPT)
• Google (Gemini)

The Challenge: US Sanctions

What makes Zhipu's achievement particularly noteworthy is the context in which it operates. Due to US sanctions, the company has:

• Limited access to top-tier Nvidia AI accelerators
• Been forced to work more closely with domestic provider Huawei Technologies
• Pivoted to using Huawei's Ascend chips for training

First Fully Domestic Chip-Trained Multimodal Model

This month, Zhipu also released a new multimodal model that it claims is China's first to be fully trained using domestic chips. The company's open-source image generation model, GLM-Image, is described as the first state-of-the-art multimodal model to complete training entirely on Huawei's Ascend chips.

A Growing Trend

Zhipu is not alone in facing overwhelming demand. Last year, Chinese startup DeepSeek experienced similar surges and was forced to restrict access to its API service to manage server capacity.

This pattern highlights:

1. Explosive growth in AI tool adoption
2. Infrastructure challenges in scaling AI services
3. The global nature of AI development, with strong contenders emerging from China

What This Means for Developers

For the developer community, this development signals:

• AI coding assistants are becoming essential tools
• Infrastructure capacity remains a key bottleneck
• Competition in the AI space is heating up globally
• Regional alternatives are emerging despite geopolitical constraints

Looking Ahead

As Zhipu works to scale its infrastructure to meet demand, this temporary limitation serves as a testament to the rapidly growing adoption of AI-powered development tools worldwide. The company's successful IPO in Hong Kong this month further validates investor confidence in Chinese AI innovations.

We'll be watching closely to see how Zhipu addresses these capacity challenges and whether their domestic chip strategy proves sustainable in the long run.

Source: The Edge Malaysia - January 21, 2026