Chinese AI models are handling more U.S. traffic

Layer3Labs says Chinese-developed models now process as much as 46% of token traffic on a major U.S.-based AI router, up from about 30% a year ago. The company says many businesses do not know which model is handling their data because AI is often embedded through vendors, routing systems and shadow AI.

Why it matters: - Layer3Labs says the shift raises data-sovereignty and compliance risk for U.S. businesses that may not know where sensitive prompts are processed. - The exposure matters most in regulated industries such as healthcare, legal and finance, where customer data can move through third-party AI systems without a clear approval trail. - A Chinese-provider endpoint can fall under China's National Intelligence Law, which can compel data disclosure to the state. - Layer3Labs says firms with high shadow AI exposure face an estimated $670,000 breach premium per incident.

What happened: - Layer3Labs said as much as 46% of token traffic on the largest U.S.-based AI model router is now routed to Chinese models. - The company said U.S. models handled roughly 70% of that traffic a year ago, and now handle closer to 30%. - Layer3Labs is an AI implementation firm in Miami that audits company AI stacks. - The company says most business owners do not choose a model directly and instead buy an AI feature whose backend can be routed, resold or white-labeled without notice.

The details: - Routing defaults can send prompts to the cheapest available provider, and Layer3Labs says Chinese inference can run 60% to 90% cheaper. - Andreessen Horowitz estimates 80% of U.S. startups build on Chinese base models, with named adopters including Airbnb, Uber and Cursor. - Research cited by Layer3Labs puts 70% of enterprise AI activity outside IT oversight, with many of those models small enough to run on a laptop and bypass security review. - Layer3Labs said its review of dozens of small- and mid-sized business AI stacks found 70% contained at least one Chinese model the company had not knowingly approved. - The firm is offering a free AI Stack Sovereignty Check to identify which models are processing a company's data and to flag data-sovereignty and compliance risk. - Layer3Labs says businesses should run a provenance check on every AI tool touching customer data. - The company was founded in 2019 and says it helps small and mid-size businesses deploy AI safely in 2 to 4 weeks. - Layer3Labs says it specializes in AI strategy, workflow automation and vendor risk assessment. - The company says learn more at Layer3Labs.

Between the lines: - The core issue is visibility, not just geography: businesses may be buying convenience while losing control over which model processes their data. - Cost pressure appears to be a major driver, since cheaper Chinese endpoints can become the default path in routing systems. - The market is shifting faster than many compliance programs, which creates a gap between AI adoption and oversight. - In July 2026, the House Select Committee on the CCP opened a probe into U.S. firms' use of these models, and the NDAA already bars DeepSeek from Defense Department systems.

What's next: - Layer3Labs is urging companies to map every AI tool that touches customer data and to confirm the model behind each product. - Regulated businesses are likely to face more pressure to document AI provenance as use of vendor-backed and shadow AI expands. - The ongoing federal scrutiny of Chinese models could push more U.S. companies to review routing settings, vendor contracts and internal AI policies.

The bottom line: - The biggest AI risk for many companies may not be what they chose, but what their vendors routed behind the scenes.