· via TechCrunch
US tightens AI chip export rules to block remote compute access from China
The Commerce Department's interim rule requires licenses for AI chips and servers shipped to any data center worldwide that could serve companies in China or other embargoed countries, replacing the Biden-era AI Diffusion Rule.

The Trump administration is preparing to roll out new export controls on AI chips and overseas data centers, according to TechCrunch. Under an interim rule from the Commerce Department, companies will need a license to export, reexport or transfer AI chips and servers to data centers anywhere in the world if those facilities may provide compute to firms in China or other restricted countries — closing off remote access as a workaround to existing chip bans.
What the rule requires
The remote-access provisions target companies based in the US "D:5" group of embargoed countries, a list that includes China, Russia, Iran, North Korea and 18 others. Existing rules already require chipmakers to obtain licenses to ship AI chips to companies based in, or owned by businesses in, those countries; the new control extends that coverage to compute consumed from abroad.
The rule itself describes the measure as "a logical extension of existing controls," arguing the restrictions will undermine the ability of companies headquartered in countries of concern to use US advanced "Super Intelligence" (SI) semiconductors to train advanced models.
Compute threshold and compliance duties
The rules apply to data centers, or any network, capable of supplying more than 12 million total processing performance (TPP) — an export-control metric for computing power that TechCrunch notes roughly equals the peak output of about 750 Nvidia H100 or H200 chips.
Indirect access arranged through brokers or intermediaries is also covered. Exporters and data center operators are expected to monitor their customers' ownership structures and activities to confirm they are not controlled by anyone from the D:5 group. Operators of existing facilities whose services fall under the rule must obtain licenses within 90 days of it taking effect.
One carve-out: commercial recommendation systems, such as those used in e-commerce or social media, are not covered.
New license requirements for five countries
The interim rule also adds licensing requirements for exports, reexports and transfers of advanced semiconductors to Malaysia, Singapore, Indonesia, Thailand and Brazil. The Commerce Department included these countries because they "present a higher-than-normal risk of diversion." Chips destined for companies based in the US are exempt from that requirement.
Nvidia scandal and industry backdrop
The move arrives days after Nvidia co-founder Jensen Huang resigned amid a scandal involving an alleged Chinese spy linked to one of the chipmaker's customers, TechCrunch reports. The US government has reportedly sought to place that customer, Megaspeed, on a special entity list, on the belief that it purchased thousands of Nvidia chips to supply compute to Chinese businesses and sidestep existing export-control restrictions.
Citing Reuters reporting from last week, TechCrunch adds that Chinese AI companies have been racing to secure computing capacity overseas in anticipation of tighter US restrictions.
The AI Diffusion Rule is rescinded
The interim rule also formally rescinds the Biden administration's AI Diffusion Rule from January 2025, which imposed country-level limits on advanced AI chip exports and restricted transfers of certain powerful AI models. The Trump administration had already said in May that it would discard that framework, arguing it placed unnecessary restrictions on US companies and allies.
The stated rationale for the new measures is national security. The rule says access to large-scale computing power can be used to train and run advanced models capable of supporting military- and intelligence-related activities that could threaten US national security or foreign policy interests.
Why it matters
Anyone building or operating AI infrastructure across borders now faces an additional layer of licensing and compliance obligations, including continuous scrutiny of who ultimately controls their customers. Overseas data center and cloud operators serving international clients will need to verify ownership structures within 90 days or risk losing access to US chips. For AI developers in China, the rule targets the workaround that direct chip bans created — renting compute abroad — and the reported rush for overseas capacity suggests demand has already shifted. The rescission of the AI Diffusion Rule also marks a policy turn away from broad country quotas toward targeted enforcement against diversion, reshaping planning for chipmakers, hyperscalers and anyone renting large GPU fleets.
- #export-controls
- #ai-chips
- #nvidia
- #data-centers
- #us-policy