• Treasury Secretary Scott Bessent said proper chip restrictions will be in place and remote access has been curtailed.
  • The Commerce Department this week unveiled a rule targeting data centers that supply advanced computing power to Chinese firms, with a 90-day transition for existing arrangements.
  • Nvidia (NASDAQ:NVDA), the primary company affected, recently reported strong results but faces uncertainty over overseas exposure.

Treasury Secretary Scott Bessent on Friday moved to reassure that the U.S. is tightening controls on advanced semiconductors, stating that proper chip restrictions will be enforced and that remote access has been curtailed. His remarks come after the Commerce Department this week announced a new rule aimed at preventing Chinese companies from bypassing export restrictions by renting computing power in overseas data centers.

A New Front in Tech Restrictions

The rule, titled “Addressing Remote Access to Advanced Semiconductors,” was released by the Bureau of Industry and Security on November 16 and targets data centers worldwide that provide substantial computing power to companies headquartered in China, Macau, or other restricted destinations. It also adds licensing requirements for advanced-chip shipments to Indonesia, Malaysia, Singapore, Thailand, and Brazil, which BIS considers higher-risk diversion routes.

A conditional exclusion exists for U.S.-headquartered recipients, coupled with monthly reporting requirements. The rule’s effective date is November 19, but existing covered arrangements have a 90-day licensing transition, meaning Bessent’s comment should not be read as confirmation that all affected access has already stopped.

The action follows scrutiny of Singapore-based Megaspeed International, which authorities suspect of diverting Nvidia chips. U.S., Singaporean, and Malaysian authorities have investigated aspects of its activities; suspected violations are not established findings of guilt.

The policy shift has also exposed an administration disagreement: Treasury and the National Security Council pushed for the restrictions, while Commerce Secretary Howard Lutnick faced criticism over the pace of action. President Trump acknowledged the tension while emphasizing the importance of the U.S.-China trade relationship.

Nvidia in the Crosshairs

Nvidia is the principal company implicated. Its products span AI accelerators, complete data-center systems, networking, software, and graphics/edge-computing platforms. On November 17, Nvidia reported fiscal third-quarter revenue of $115.8 billion, up 103% year over year, with data-center revenue of $107.6 billion, up 110%. GAAP net income was $65.3 billion, or $2.70 per diluted share.

For the current quarter, management forecast revenue of $130 billion, plus or minus 3%, explicitly assuming no data-center compute revenue from China. That guidance limits the usefulness of treating the new rule as a simple reduction to direct China sales; the additional exposure concerns overseas customers serving Chinese workloads and the compliance risks around those relationships.

In a separate development, AP reported that Jensen Huang resigned as CEO last week amid scrutiny connected to a former Megaspeed executive. Nvidia’s earnings release identifies Colette Kress as interim CEO while retaining her executive vice president and CFO titles. The circumstances reported by AP should remain distinct from any proven legal wrongdoing.

Nvidia also announced an agreement to acquire Hugging Face and said its Vera Rubin NVL72 platform had begun shipping, signaling that its broader product and expansion agenda continues alongside the regulatory disruption.

Compliance Costs and Global Tensions

The new end-use control generally applies above 12 million aggregate total-processing-performance units supplied to covered companies. Linked data centers can count together, and indirect arrangements are included. Being below that threshold does not automatically make an arrangement permissible, as other existing controls may still apply.

The main economic effects are likely to include higher compliance costs for chip suppliers and cloud operators, including ownership checks, customer screening, contractual information rights, and ongoing monitoring. Data-center projects in the five newly designated countries face greater uncertainty, as licensing could delay equipment deliveries or require changes to customer mixes.

China has objected to the measures. Foreign Ministry spokesperson Guo Jiakun called them economic coercion and “long-arm jurisdiction,” arguing that they interfere with legitimate trade and global supply chains. Beijing said it would take necessary steps to protect its interests, but no specific retaliatory measure has been established.

The timing is diplomatically sensitive: the announcement coincided with Trump’s departure for Asia and anticipated talks with Xi Jinping amid a fragile trade truce. It therefore tests whether Washington and Beijing can separate commercial negotiations from restrictions on strategic technology.

The policy debate centers on whether stronger controls successfully prevent circumvention or impose excessive costs on legitimate international commerce. U.S. lawmakers have argued that cloud access undermines physical-chip restrictions; China argues that the expanded controls disrupt lawful technological cooperation.

For Nvidia, the key uncertainty is how much overseas demand depends on workloads now requiring authorization. Its latest guidance excludes direct China data-center compute revenue, but neither that guidance nor Bessent’s statement quantifies the broader overseas exposure. The evidence therefore supports a tightening of access and heightened enforcement risk—not a verified complete cutoff or a dependable estimate of the resulting revenue loss.