China called Anthropic's Claude Code a backdoor and banned it, while the US restricts Chinese models

The US-China confrontation in artificial intelligence escalated sharply in early July 2026: Beijing officially declared Claude Code from Anthropic a backdoored tool and urged companies to abandon it, while American lawmakers simultaneously initiated restrictions against firms dependent on cheap Chinese AI models.

What exactly was found in Claude Code?

China's Ministry of Industry and Information Technology (MIIT) announced that its official cybersecurity platform detected Claude Code — Anthropic's AI assistant for code writing — as an information security threat. Chinese organizations were recommended to cease using the tool immediately.

The MIIT did not disclose specific technical details of the accusation. Nevertheless, the official status of the statement carries real consequences: state-owned companies, banks, and organizations working with sensitive data must respond without delay. Ignoring recommendations from a regulator of this level carries legal and reputational risks — regardless of how technically justified the accusation is.

Key facts:

• Claude Code is an AI assistant for developers from Anthropic, one of the leading US AI laboratories
• The accusation was raised by an official cybersecurity platform under China's MIIT
• Chinese companies were recommended to discontinue use of the product
• Anthropic has not provided public comments as of publication

Claude Code ranks among the most widely used AI tools in the developer community: it enables writing, debugging, and documenting code through dialogue with a language model. Its widespread adoption in China's IT segment makes the ban noticeable — affecting both international companies with offices in China and local developers who have integrated the tool into daily workflows.

Why are the US restricting Chinese models?

Simultaneously with the Chinese ban, American lawmakers are advancing measures against US companies building products on top of cheap Chinese AI models. The rationale is symmetrical: dependence on infrastructure from a potentially hostile state creates systemic risk to national security, even if it concerns commercial software products.

Chinese models — particularly DeepSeek and Alibaba solutions — have secured significant positions in American startups and the research community thanks to a combination of low inference costs and high technical quality. The proposed legislation aims to create a barrier to this dependence before it becomes structural.

Notably, both legislative initiatives emerged practically simultaneously — within a single week of early July 2026. This is not one side's reaction to the other's actions, but a parallel movement demonstrating that the AI market is entering a new phase of structural division.

For companies operating in both jurisdictions, this means a fundamental choice: either maintain two separate AI stacks — Western and Chinese — or choose one side and accept inevitable losses in the other market.

What this means

The split in AI space along national lines is moving from the chip level to the software level. Restrictions now apply not only to hardware components, but to specific development tools — code assistants, API access to models, cloud AI platforms. Mutual accusations of "backdoors" and "unsafe dependencies" risk becoming a standard tool of technology policy. For developers and companies, this is no longer abstract geopolitics — it is a concrete question about which tools will be available tomorrow.