Calder & Vance International Sanctions & Compliance Counsel

Export Controls & Dual-Use · EU

Cloud and remote-access technology controls under EU: compliance counsel

A cloud-services provider based in the EU wins a contract to supply advanced computing resources to a research institution in a third country. Access will be delivered remotely – no hardware crosses a border. The compliance team assumes export-control rules do not apply. They are wrong. That assumption, repeated across dozens of similar arrangements every year, is precisely how companies accumulate material export-control exposure without realising it.

Cloud and remote-access technology controls under EU export-control rules capture the supply of controlled technology and software through intangible means, including remote access to computing systems, cloud environments, and software-defined infrastructure. The governing instrument is EU Regulation 2021/821 (the recast Dual-Use Regulation), administered by Member State competent authorities. A transfer of controlled technology via remote access is treated as an export for classification and licensing purposes, regardless of whether any physical item crosses a border.

This page sets out what falls within scope, how the classification and licensing process works, where the EU position diverges from the US and UK regimes, and what a business should do before it provisions access.

What falls within scope of EU cloud and remote-access technology controls?

The EU recast Dual-Use Regulation defines "export" to include the transmission of software and technology by electronic media, fax, or telephone. Remote access to a controlled system – including access to a server, a virtual machine, or a cloud environment running controlled software – is caught on that basis. The controlling question is not how the access is delivered but what the recipient can do with it.

Items that attract controls in this context include cryptographic software, certain high-performance computing resources, and technology for the development or production of items on the EU Common Military List or the dual-use annexe. If a foreign person, seated outside the EU, can interact with controlled technology hosted on EU infrastructure, that interaction is an export in regulatory terms. The seat of the server is not the determinative factor. What matters is where the recipient is located and what classification the accessed technology carries.

We regularly advise clients who operate multi-tenant cloud environments and who have assumed that physical-export controls do not reach their service model. In almost every case, at least one service tier involves controlled cryptographic or computing technology that brings the arrangement squarely within the regulation's scope. The practical question is always the same: has every service component been classified, and has every customer geography been assessed against that classification?

How does the classification process work for cloud-hosted and remote-access services?

Classification under the EU Dual-Use Regulation follows the dual-use list contained in Annex I to the regulation. Each item – hardware component, software module, or technology parameter – is mapped against the list entries, which are organised by category (nuclear, special materials, electronics, computers, telecommunications, sensors, lasers, navigation, aerospace, and maritime) and product group (equipment, test and production, material, software, technology).

For a cloud or remote-access service, classification is not a single exercise. It applies at each functional layer: the infrastructure layer (servers, networking hardware), the platform layer (operating software, virtualisation tools, encryption modules), and the service layer (the application the customer actually uses). A service may include multiple controlled items at different layers, each requiring independent classification. Treating the service as a single undifferentiated product and classifying it once is a common and expensive mistake.

Where a component does not appear on the dual-use list, it may still be subject to catch-all controls if the exporter has knowledge or grounds to suspect that it will be used in connection with weapons of mass destruction programmes, military end-uses in specified destinations, or embargoed end-users. The catch-all does not require a listing; it requires an assessment of the end-use and end-user.

Member State competent authorities – the relevant national authority in the exporter's Member State – are the point of contact for classification queries. Most Member States operate a binding-classification-opinion mechanism, though the process and timeline differ by jurisdiction. In our experience, exporters that engage early with the competent authority reduce the risk of a post-contract compliance failure significantly.

What authorisation routes are available, and how do they work?

Once a controlled item is identified, the exporter must determine whether an authorisation applies. The EU Dual-Use Regulation provides several authorisation types.

EU General Export Authorisations (EUGEAs) are standing authorisations that permit exports of specified items to a defined list of destinations without a case-by-case application. They carry conditions – on end-use declarations, due-diligence checks, and record-keeping – and are not available for all items or all destinations. A cloud provider supplying access to countries not covered by a EUGEA must assess whether a national general authorisation in its Member State applies, or whether it needs a specific licence.

A specific (individual) export licence is required where no general authorisation covers the transaction. The application is submitted to the competent authority in the exporter's Member State, and it must include technical specifications, end-user information, and a statement of intended end-use. Timelines are not fixed at the EU level; they differ by Member State and by the sensitivity of the item. In practical terms, a specific licence application for a controlled computing or encryption service should be planned well in advance of the contract execution date. Waiting until after signature – or, worse, after access has been provisioned – sharply narrows the options available.

For cloud services that involve a flow of technology across multiple EU Member States before it reaches the end-user, questions about which competent authority has jurisdiction can arise. The general position is that the exporter in the Member State from which the technology exits the EU customs territory is responsible, but multi-hop architectures can complicate that analysis.

How does EU cross-border cloud-access control differ from the US and UK regimes?

The cross-border comparison matters for any business operating under more than one regime – and most cloud and managed-service providers do.

Under the US Export Administration Regulations (EAR), a deemed export occurs when controlled technology is released to a foreign national within the United States. The concept is explicit in the EAR and drives significant compliance obligations for cloud operators, research institutions, and any employer with a mixed-nationality workforce using controlled systems. The EU Dual-Use Regulation does not have a directly equivalent deemed-export provision, though member-state implementation varies and some national instruments apply analogous controls on technology disclosure within borders. That divergence creates risk for US-EU joint ventures and shared-infrastructure arrangements: controls that apply in one regime may not apply identically in the other, but conduct that satisfies the looser standard does not necessarily satisfy the stricter one.

Foreign Direct Product rules under the EAR can extend US jurisdiction to foreign-produced items that are the direct product of controlled US technology or software. A cloud service operating on infrastructure that incorporates US-origin controlled components may bring otherwise non-US items within the EAR's reach. We advise clients on this interaction regularly, because it is one of the most overlooked points of extraterritorial exposure in the cloud sector. The EU Dual-Use Regulation does not operate extraterritorially in the same way, but EU-incorporated entities running cloud infrastructure remain subject to it regardless of where the physical hardware sits.

Under the UK export-control regime, administered by the ECJU, similar controls on intangible technology transfers apply under the Export Control Order and the relevant implementing measures. The UK and EU regimes diverged after 2020; the dual-use lists are no longer identical, and a transaction that qualifies for a general authorisation under one regime may not qualify under the other. For a business provisioning cloud access from both a UK and an EU entity to the same end-user, separate licence assessments are required.

The practical consequence of all this is that compliance for cross-border cloud services is not a single-regime exercise. Where two regimes both apply, the stricter prohibition governs the conduct of a person subject to both. The analysis must be run in parallel, not sequentially.

What are the principal risk flags in cloud and remote-access deployments?

A mis-classified item is the most direct source of enforcement exposure. If a controlled software component is treated as uncontrolled – because it is commercially available, because it appears ordinary, or because the classification process was not applied at the relevant service layer – every instance of provisioning access to a restricted destination or end-user constitutes a potential violation. Commercial availability does not, by itself, remove dual-use controls. That is one of the most persistent myths in this sector, and it is wrong.

The following patterns, drawn from our cross-border practice, generate the highest frequency of compliance failure in cloud and remote-access contexts.

  • Layer-by-layer classification gaps: the service is treated as a single product, and the cryptographic or high-performance computing components at lower infrastructure layers are never individually assessed.
  • End-user diligence that stops at the contractual customer: the immediate customer is a low-risk entity, but the technology is sub-provisioned or shared with a higher-risk end-user down the chain.
  • Geography drift in multi-region deployments: access is initially provisioned to a compliant destination, but the customer subsequently routes usage through a node in a restricted jurisdiction. Monitoring obligations do not end at the point of provisioning.
  • Catch-all triggers not systematically assessed: the item is not on the dual-use list, so no classification is run, and the catch-all obligation – which requires an assessment of end-use and end-user regardless of list status – is never applied.
  • Record-keeping gaps: transaction records, due-diligence documentation, and end-user undertakings are not retained for the required period, making enforcement defence substantially harder.

Are your end-use monitoring procedures designed for a physical-goods environment and simply assumed to apply to cloud provisioning? That assumption produces the geography-drift risk almost every time.

When should a business involve specialist export-control counsel?

The position above covers the standard case. Your facts – the technology stack, the customer base, the Member State of export, and the regimes in play – change the analysis. Specialist EU export-controls counsel should be involved at the following points.

Before a new service is launched, or a new geographic market is opened, a pre-launch classification review confirms which service layers are controlled, which authorisations are available, and what conditions attach to them. This review is most valuable before contract execution; it is significantly harder and more expensive to unwind a provisioned service that turns out to be unlicensed.

When a compliance team identifies a transaction that may not have been properly licensed, the priority is to scope the exposure, assess whether voluntary self-disclosure is appropriate, and preserve the records that will be needed for any regulatory engagement. Voluntary self-disclosure (a recognised mitigation route under most regimes) is most effective when it is made promptly and with a complete account of the facts. Delay does not improve the position.

When a customer's end-use or end-user status changes – through an acquisition, a restructuring, or a shift in the customer's own customer base – an updated assessment is needed. Export-control authorisations are granted on the facts presented at the time of application; material changes to those facts can affect the validity of the authorisation.

If a transaction has already been flagged, or a filing has been refused, an early review can preserve options that narrow with time.

How Calder & Vance supports cloud and managed-service providers

We act for multinationals, financial institutions, and exporters on sanctions and export-control matters across the major regimes. In the cloud and remote-access sector, our work centres on four areas.

Classification and licensing: we assess eligibility, prepare and submit licence applications across EU Member State competent authorities, and manage the regulator's queries through to determination. Where a EUGEA or national general authorisation applies, we identify it and confirm the conditions that must be met to rely on it.

Programme design: we test the classification logic applied to a client's service architecture, identify the layers where controls have not been applied, and redesign the compliance programme to address them. This includes end-user diligence procedures, usage-monitoring controls, and the record-keeping architecture needed to support enforcement defence.

Cross-regime analysis: where a client operates under the EAR as well as the EU Dual-Use Regulation, or where the UK regime is also in play, we run parallel assessments and identify the points of divergence that require separate treatment. We engage local counsel in relevant jurisdictions where national implementation creates specific issues.

Enforcement and voluntary self-disclosure: where a potential breach has been identified, we scope the apparent violation, advise on voluntary self-disclosure, and prepare the penalty defence. In our experience, the quality of the initial response to an enforcement inquiry shapes the outcome more than almost any other variable.

To discuss a classification review or a licence application under the EU Dual-Use Regulation, or to stress-test your export-control programme for cloud and remote-access services, contact Calder & Vance at info@caldervance.com.

Related practices

Frequently asked questions

How long does control remote access to computing power take under EU?
There is no fixed EU-level timeline for a specific export-licence application covering remote access to controlled computing resources. Processing time is set by the competent authority in the relevant Member State and depends on the sensitivity of the technology, the destination, and the quality of the information provided with the application. In practice, applications for controlled high-performance or cryptographic computing services should be planned well ahead of contract execution – weeks to months is the realistic range, not days. A binding classification opinion, where available, should be sought even earlier. Attempting to licence after access has already been provisioned reduces the options materially and may itself constitute a reportable breach.
What are the main risks in cloud and remote-access technology controls under EU?
The primary risks are: classification gaps at the infrastructure or platform layer, where controlled cryptographic or computing components are not individually assessed; catch-all failures, where an unlist item is provisioned to a high-risk end-user without an end-use assessment; geography drift in multi-region deployments, where usage migrates to a restricted jurisdiction after provisioning; and inadequate end-user diligence that stops at the contractual customer rather than following the technology to its ultimate user. Each of these can produce a violation even where the initial contracting decision appeared sound. Record-keeping failures compound the risk at the enforcement stage, making it substantially harder to demonstrate that a due-diligence process was applied.
Do we need specialist counsel for cloud and remote-access technology controls?
For most cloud and managed-service providers operating cross-border under the EU Dual-Use Regulation, specialist export-control counsel is not optional – it is the practical means by which classification accuracy is achieved and licence applications are prepared to the standard competent authorities expect. The regulation is technically demanding at every layer: the dual-use list, the catch-all controls, the end-use and end-user assessment, and the interaction with the US EAR and UK export-control regime each require expertise that a general compliance function is unlikely to hold in-house. Businesses that engage counsel before a service launches or a new market is opened consistently achieve better outcomes than those that engage after a regulatory inquiry has commenced. Law firm for cloud and remote-access technology controls EU matters should be selected on the basis of demonstrated cross-regime capability, not regime-specific silo expertise.

Talk to Calder & Vance

For a scoped view of your exposure, contact info@caldervance.com.

Discuss your matter

Discuss this with Calder & Vance

A few lines on your situation and any deadline. We respond within two hours in business hours.

This publication is general information and does not constitute legal advice. For advice on your situation, contact info@caldervance.com.