What The US Remote Access Rule Means For Your Cloud

A new US export rule requires licences for AI chips in data centres anywhere that serve companies headquartered in China. Here is what the remote access rule means for European cloud customers, their contracts and their ownership checks.
AI generated image - a quiet European data hall with server racks and a paper customer file on a cart, illustrating the US remote access rule for cloud providers

A US export rule now reaches data centres in Frankfurt, Dublin and Amsterdam, through one question: who are their customers, and who owns those customers?

The US Commerce Department’s Bureau of Industry and Security (BIS) has issued the remote access rule. It is an interim final rule called “Addressing Remote Access to Advanced Semiconductors”, and BIS expects it to appear in the Federal Register on 19 November 2026, the day it takes effect. For a European organisation that buys cloud or colocation, the questions it raises are practical ones, and they arrive through your providers.

What the remote access rule requires

The remote access rule targets computing power, not only chips in boxes. According to the client alerts of Pillsbury and WilmerHale, a licence is now required to export, re-export or transfer advanced AI chips and servers to a data centre anywhere in the world when that data centre gives more than 12 million “aggregate TPP” of computing power to companies headquartered in China, Macau or the other countries in US Country Group D:5. TPP is a measure of chip performance. The threshold also counts when a company’s ultimate parent sits in one of those countries.

The count is not limited to one building. It runs across “linked data centers”, and Nixon Peabody notes that this includes sites linked digitally, not only physically. Data centres that already serve covered companies at this level get 90 days to obtain a licence. Shipments already on their way may go ahead if they arrive within 30 days.

The remote access rule also adds a licence requirement for advanced chips going to Indonesia, Malaysia, Singapore, Thailand and Brazil. Shipments to companies headquartered in the US are excluded, as long as the exporter reports them to BIS every month. At the same time, BIS formally rescinded the AI Diffusion Rule of January 2025, with its country tiers and its controls on model weights.

China’s commerce ministry responded to the remote access rule on 17 November. “If the US persists, China will take countermeasures to firmly safeguard its sovereignty, security and development interests,” it said, as The Next Web reported. What those countermeasures would be, the ministry did not say.

Why a European data centre is in scope

None of the three law firm alerts mentions an exemption for EU countries. The remote access rule applies to exporters, re-exporters and anyone who transfers the hardware within a country. A cloud provider in Germany that buys new GPU servers from a US-origin supply chain is a recipient of covered items. If it serves a customer whose parent company is in China above the threshold, the chips it uses for that customer need a licence.

The part that will reach you is the knowledge standard. Nixon Peabody explains that an exporter can only show it did not know where the compute went if it obtains the data centre’s know-your-customer checks, including who owns and controls its customers. A data centre that refuses or cannot provide that information is treated as a red flag. That presumption gives the exporter knowledge unless the flag is resolved.

So the chip vendor asks the data centre. The data centre then has every reason to ask you.

What the remote access rule means for your contracts

Expect three things from your cloud and colocation providers in the coming months.

The first is an ownership question. Who is your ultimate parent, and where is it headquartered? A provider that cannot answer that about its customers puts its own supply of chips at risk. For most European organisations the answer is easy. For a group with a parent outside the EU, or with a shareholder structure that runs through several holding companies, it may take longer than the provider wants to wait.

The second is a contract clause. Providers will want the right to restrict or end a service when serving a customer would breach US export controls or put their licence at risk. Some standard terms already contain a general export control clause. After the remote access rule, read it as an operational clause rather than legal boilerplate, because it now describes a situation that can actually happen.

The third is capacity. The new licence requirement covers Singapore, Malaysia, Thailand, Indonesia and Brazil. If your workloads, a disaster recovery site or an AI vendor’s inference run in a cloud region in one of those countries, ask the provider whether new hardware there is now subject to a licence and what that means for their expansion plans. We looked at why capacity has become political in Why data centre capacity just became a political question.

The question your due diligence probably skips

Vendor due diligence usually asks where data is stored and who can access it. It rarely asks who else uses the same compute. Under the remote access rule, that second question matters for continuity. If a provider loses its licence for one site because of another customer, your service on that site may be affected too.

This is the same pattern we described in Compute Sovereignty Is Not About Where The Building Is. A data centre inside the EU can still be governed by decisions made elsewhere. The remote access rule makes that concrete. A European address does not take a European data centre outside US export controls.

It also extends a point from Your Export Control Check Probably Ends At Delivery. Export compliance used to stop when the hardware arrived. Now it follows the computing power to the end user, and the end user may be a company you share a rack with.

Three things to do before the 90 days run out

None of this calls for panic. It does call for an hour with procurement and legal.

  1. Check your own ownership chain. If your group has a parent or a significant shareholder outside the EU, find out where the ultimate parent is headquartered, and have the answer ready before a provider asks.
  2. Read the export control clause in your main cloud and colocation contracts. What can the provider suspend and with how much notice? What happens to your data if it does?
  3. Ask your main AI and cloud vendors two questions. Do any of your regions depend on hardware that now needs a US licence? And has the remote access rule changed your plans for capacity in Europe?

If the answers are vague, that is useful too. It tells you where your AI supply chain depends on decisions you do not see.

BIS takes comments on the remote access rule for 60 days after publication. It is an interim rule, so the details may still move.

If you want a monthly view of rules like this one, Prep Report Trends is free on futureprep.eu.

Bas Hennis

Future Prep helps organizations prepare for the impact of AI and emerging technologies. We provide hands-on training, strategic advice, and smart tools for the responsible use of AI, governance, and digital resilience.LinkedIn

Newsletter
[sibwp_form id="1"]
Related Blogs
LATEST NEWS

AI governance is not a future problem

Regulation is already in effect. Your competitors are already building internal capability. The gap between ‘we are aware of AI’ and ‘we have operational control’ is closing, and it closes faster with a structured framework.

 

Book a 30-minute discovery call. No obligation. We will assess where your organisation stands and what a realistic starting point looks like.

No sales pressure. No jargon. Just a structured conversation about your organisation's AI readiness.

Scroll to Top