Chinese AI Models Now Handle Nearly Half of U.S. Router Traffic

A growing share of U.S. business AI traffic is being routed to Chinese-developed models, often without companies realizing it. Layer3Labs says the shift raises data-sovereignty and compliance risks as vendors, routers and shadow AI obscure where customer data is processed.

Why it matters: - Up to 46% of token traffic on a major U.S.-based AI model router now goes to Chinese models. - Layer3Labs says most business owners do not know which language model is reading their data. - The exposure matters for regulated industries such as healthcare, legal and finance, where model provenance and data location can affect compliance. - Any prompt handled by a Chinese-provider endpoint can fall under China’s National Intelligence Law, which can compel data disclosure. - Companies with high shadow AI exposure face an estimated $670,000 breach premium per incident.

What happened: - Layer3Labs, an AI implementation firm that audits company AI stacks, said Chinese-developed models now process a fast-growing share of American business AI traffic. - A year ago, U.S. models held roughly 70% of that traffic; today the share is closer to 30%. - Layer3Labs tied the shift to hidden routing inside business software, vendor backends and unsanctioned employee use. - In July 2026, the House Select Committee on the CCP opened a probe into U.S. firms’ use of these models. - The NDAA already bars DeepSeek from Defense Department systems.

The details: - Cost-optimizing routers increasingly send prompts to the cheapest available provider, and Layer3Labs says that often means a Chinese endpoint where inference runs 60% to 90% cheaper. - Andreessen Horowitz estimates 80% of U.S. startups build on Chinese base models. - Named adopters cited in the source material include Airbnb, Uber and Cursor. - Research puts 70% of enterprise AI activity outside IT oversight, much of it using Chinese models small enough to run on a laptop. - 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. - Layer3Labs is offering a free AI Stack Sovereignty Check to identify which models are actually processing company data and flag data-sovereignty and compliance risk. - Layer3Labs says the service is aimed especially at businesses handling customer data. - The company says its AI implementation and automation work for small and mid-size businesses is typically live in 2 to 4 weeks. - Layer3Labs was founded in 2019 and is based in Miami.

Between the lines: - The core problem is visibility, not just vendor choice, because businesses often buy an AI feature without knowing the model underneath. - White-labeling, resale and router defaults can move data to a different model than the buyer expects. - Shadow AI creates a second risk layer because employees can use tools that bypass security review entirely. - The trend suggests AI supply chains are becoming a procurement and governance issue, not just a technical one.

What's next: - Layer3Labs is urging businesses to run a provenance check on every AI tool that touches customer data. - The firm is especially pressing regulated companies to identify where prompts are routed and under which jurisdiction the data may fall. - Expect more scrutiny from policymakers and corporate security teams as Chinese-model usage becomes harder to ignore.

The bottom line: - U.S. businesses may be relying on Chinese AI models far more than they realize, and the hidden routing of sensitive data is now the bigger risk.