On November 16, 2026, the U.S. Commerce Department's Bureau of Industry and Security (BIS) released an interim final rule that closes one of the biggest gaps in America's AI chip controls: renting advanced U.S. computing power from data centers in third countries. The rule matters far beyond Washington and Beijing. Cloud GPU pricing and availability across Asia, including the capacity Indian startups and enterprises rely on, sit downstream of these decisions. Here is what the rule actually says and what it could mean for businesses in India.
What the rule does
The rule, titled "Addressing Remote Access to Advanced Semiconductors," is designed, in BIS's own words, "to ensure that existing restrictions on access to sensitive American technology cannot be circumvented through remote computing arrangements involving data centers in third countries." Until now, a company barred from buying advanced Nvidia chips could often rent the same computing power from a data center in Southeast Asia or elsewhere. The new rule treats that remote access much like a physical shipment.
The key mechanics
According to an analysis by law firm Nixon Peabody, the core changes are:
- A new data center end-use control. A license is now required to export, re-export or transfer advanced chips and servers when the exporter knows they are destined for a data center anywhere in the world providing computing above a defined threshold to companies headquartered in China, Macau or other restricted destinations. Capacity across digitally "linked data centers" counts toward the threshold.
- Know-your-customer diligence. Exporters are expected to obtain information about data centers' customer vetting; a refusal to share it is treated as a red flag.
- A 90-day clock for existing arrangements. Data centers already serving restricted customers must obtain a license within 90 days of the rule taking effect or face future restrictions.
- New license requirements for five diversion-risk countries: Indonesia, Malaysia, Singapore, Thailand and Brazil, with case-by-case review.
- The Biden-era AI Diffusion Rule is formally rescinded, including its model-weights controls and country tiers, and the regulation adopts the term "super intelligence" in place of "artificial intelligence."
The rule takes effect on publication in the Federal Register, expected November 19, with public comments due 60 days later.
Why now
The immediate context is enforcement. U.S. officials have grown concerned that export controls were being bypassed through overseas compute rental, and recent scrutiny of chip diversion through Southeast Asian data centers gave the issue urgency. Notably, BIS says the rule was written by and at the direction of the Treasury Department and the National Security Council, an unusual interagency footprint that signals more chip-related controls may follow.
A quick note on the rescinded rule
The AI Diffusion Rule being removed was the outgoing administration's framework that sorted countries into tiers and capped how many advanced chips each could import. Its repeal does not mean open access: the new rule replaces broad country tiers with targeted controls aimed at specific end uses and diversion routes. For businesses, the practical takeaway is that the compliance burden is shifting from "where the chips are" to "who ultimately uses the compute."
What it means for Indian businesses
India is not on the restricted list or the five diversion-risk countries, so the direct legal exposure for Indian companies is limited. The practical effects are indirect but real. Many Indian AI startups and enterprises rent GPU capacity from cloud regions in Singapore and Malaysia, whose operators now face new licensing and due-diligence duties; some of that compliance cost and friction could show up in Asian GPU cloud pricing and availability. At the same time, tighter controls on third-country data centers make India's own data-center build-out more attractive, both as a market and as an alternative location for compliant compute. If your AI roadmap depends on rented GPU capacity, this is a good moment to ask your provider where your workloads physically run and how exposed that capacity is to U.S. licensing shifts.
What to watch
Three things over the coming weeks: the Federal Register publication, which starts the 90-day and comment-period clocks; any additional country designations, since the rule lets the Commerce Secretary add more; and guidance on the narrow carve-out for commercial recommender systems. We will keep tracking the story.
If your business is planning AI workloads or cloud infrastructure and wants help thinking through resilience and cost, VAIR IT offers a free consultation for businesses in Delhi NCR and across India.



