AI theft and cyber mercenaries collide: US targets crypto scams as China fires back

On September 9, 2026, China publicly rejected U.S. claims that Chinese actors are conducting “industrial-scale” AI theft, responding to accusations by three U.S. agencies that Chinese AI companies exploited American AI models. In parallel, U.S. lawmakers urged the Treasury Department to sanction three India-based hackers-for-hire groups accused of stealing data from thousands of U.S. citizens and companies, framing the issue as a mercenary cyber-economy that evades accountability. The U.S. Department of Justice also announced coordinated disruption of an illicit online marketplace called Xinbi Guarantee, including seizure of Telegram channels used to run the service and freezing $52.8 million in cryptocurrency tied to scam operations. Separately, threat researchers reported that multiple spy groups used the same Chrome and Windows exploit kit within a week, pointing to a China-aligned cluster deploying a previously undocumented “BlueMoon” chain of vulnerabilities across Microsoft Windows and Google Chrome. Strategically, the cluster shows a widening U.S.-China cyber and AI contest that is moving from accusations into operational disruption and potential sanctions. China’s rebuttal suggests Beijing is trying to contest the evidentiary and political framing of “theft,” while Washington is simultaneously escalating enforcement through DOJ actions and legislative pressure on Treasury. The lawmakers’ push to sanction India-based mercenaries adds a third-country enforcement layer, implying that attribution and deterrence may increasingly target cyber supply chains rather than only state-linked operators. The exploit-kit reporting reinforces that espionage tradecraft is being standardized and reused quickly, which raises the risk that AI model theft claims are part of a broader intelligence and capability-accumulation strategy. Market and economic implications are most visible in cybersecurity risk premia, insurance and incident-response demand, and the compliance posture of firms handling sensitive data. Sanctions and enforcement against hackers-for-hire can tighten the illicit cyber-services ecosystem, potentially affecting the cost and availability of downstream “data monetization” and fraud infrastructure. The DOJ’s $52.8 million crypto freeze is a concrete signal for crypto-linked scam liquidity, which can influence exchange risk controls and the behavior of scam-adjacent token flows, even if the broader crypto market impact is likely limited. For technology markets, the reported BlueMoon exploit kit and the AI theft dispute can raise near-term volatility in enterprise software and cloud security spending, while also increasing scrutiny of AI supply chains, model licensing, and vendor risk management. What to watch next is whether Treasury follows the lawmakers’ request with sanctions designations and whether additional indictments or asset freezes expand beyond Xinbi Guarantee. In the cyber domain, monitor indicators of BlueMoon reuse—new victimology, additional vulnerability chaining, and whether Microsoft or Google issues rapid mitigations that reduce exploitability. On the AI front, track whether the U.S. agencies behind the “industrial-scale” claims publish more technical evidence or move toward export controls, procurement restrictions, or model-access limitations. A key trigger for escalation would be any follow-on claim that stolen AI models are being integrated into U.S. critical systems or major commercial products, while de-escalation would look like evidence-sharing mechanisms, joint incident reporting, or narrower enforcement focused on non-state criminal infrastructure rather than broad state attribution.

Geopolitical Implications

• 01

  The U.S.-China AI narrative is hardening into a security enforcement cycle, increasing the likelihood of export-control or procurement constraints even without kinetic conflict.

• 02

  Third-country targeting (India-based mercenaries) indicates a shift toward dismantling cyber supply chains and deterrence through sanctions rather than only attribution.

• 03

  Standardized exploit-kit chaining across major consumer and enterprise platforms raises the probability of cross-sector espionage and broader cyber spillover.

Key Signals

• —Treasury sanctions designations: whether the requested India-based groups are formally named and what evidence package is cited.
• —Additional DOJ actions: follow-on indictments or asset freezes expanding beyond Xinbi Guarantee and its Telegram infrastructure.
• —Vendor patch cadence: Microsoft and Google mitigation releases that reduce BlueMoon exploitability and whether attackers pivot to new chains.
• —AI governance moves: any U.S. export-control, model-access, or procurement restrictions tied to the “AI theft” allegations.