On 15 November 2023 President Xi Jinping and President Biden met in San Francisco. They agreed to restore and establish communication channels, manage differences and advance cooperation, including the launch of an intergovernmental dialogue on artificial intelligence. It was an important stabilising signal after a period of tension.
For businesses, two layers must be distinguished: political communication and regulatory enforcement. The former improves expectations; the latter continues under each jurisdiction’s own law. Conflating them leads to poor compliance judgments.
What the meeting delivered
The outcomes centred on communication mechanisms: resuming military-to-military channels, cooperation on narcotics, restarting climate dialogue, promoting people-to-people exchanges and agreeing to discuss AI governance. These arrangements reduce the risk of miscalculation and create channels for specific negotiations.
The meeting did not change export controls, investment screening, tariffs or entity lists. Those measures follow their own legal procedures and do not adjust automatically after a summit.
Three effects on cross-border businesses
The first is expectations. Restored communication lowers the probability of abrupt policy shifts, allowing companies to carry less political-risk premium in long-term investment and supply-chain planning. Improved expectations are not the disappearance of risk.
The second is compliance. Export controls, outbound investment screening, cross-border data, anti-money laundering and sanctions remain separate obligations. Companies must assess counterparty, item, destination and technology type one by one, rather than treating improved relations as an exemption.
The third is contracts. Geopolitical uncertainty should be allocated in contracts: termination rights on policy change, treatment of force majeure and change of circumstances, alternatives for supply-chain disruption, and compliance representations and warranties.
What the AI dialogue means
The agreement to launch an intergovernmental AI dialogue shows that AI governance has become a standing item between the two governments. It creates no immediate filing obligation, but it will shape future coordination.
If the two sides develop common language on model safety, risk classification and testing methods, cross-border compliance materials may become easier for both regulators to understand. Differences remain, especially where data export, algorithmic transparency and export controls intersect.
Companies should prepare two sets of materials: security assessment and algorithm filing documents for Chinese regulators, and model descriptions, risk reports and export-control classifications for the United States and other jurisdictions. The two can share a factual base, but one document cannot be assumed to serve both.
Four steps for companies
First, map cross-border transactions by item, technology, software and service, and assess exposure to export controls, investment screening and sanctions lists.
Second, update contract clauses to allocate policy change, licence withdrawal and supply-chain disruption explicitly.
Third, build dual compliance files: one for China, one for other jurisdictions, factually consistent but separately concluded.
Fourth, monitor implementation. Post-summit changes usually appear in departmental notices, list updates and enforcement cases, not in leaders’ statements.
For multi-jurisdiction assessments, see our multi-jurisdiction service; for export controls and item classification, see export controls; for cross-border tax and structuring, see overseas tax.
This article is based on public reporting of the Xi-Biden meeting in San Francisco on 15 November 2023. It is for reference only and does not constitute legal advice.