Does the EU AI Act Apply to US Companies? Territorial Scope Explained
The Act does not care where you are incorporated. Four hooks in Article 2 decide whether a US company sits inside EU AI Act scope — and two of them are about your customers, not you.
The first question every US vendor asks is “does this apply to us?” — and the Act answers it before you reach the obligations. Article 2 defines scope by what you do and where your output lands, not by your place of incorporation. A Delaware C-corporation with zero EU staff can be fully in scope; an EU subsidiary using a US model can be the deployer while the US parent is the provider.
This matters commercially, not just legally. EU enterprise buyers now ask for your AI Act position in security questionnaires, and the answer determines whether the deal stalls.
The four hooks in Article 2
| Hook | Article | Trigger | What it pulls in |
|---|---|---|---|
| Provider placing on the Union market | Art. 2(1)(a) | You build or brand an AI system and make it available in the EU — from anywhere | Full provider duties: conformity, documentation, registration |
| Deployer established in the Union | Art. 2(1)(b) | Any EU entity puts your system into service for its own purposes | Deployer duties fall on the EU entity, not you — but your contracts and evidence must support them |
| Output used in the Union | Art. 2(1)(c) | Your system’s output is used in the EU, regardless of where you are established or where the user sits | Provider (and for models, GPAI) duties travel with the output |
| Importer or distributor | Art. 2(1)(d) | You place an AI-enabled product on the EU market that someone else built | Importer verification duties |
Hook (c) is the one that catches US companies. If your model’s output is consumed inside the Union — an EU customer calls your API, your recommendation engine prices a European shopper, your hiring tool screens an EU candidate — you are in scope. The European Commission drew this parallel explicitly when the Act was negotiated: scope follows the GDPR Article 3 logic, where serving EU users is enough.
What scope actually obliges you to do
Being in scope is not a compliance program in itself; it determines which duties attach.
- High-risk system provider? You must appoint an authorised representative established in the Union before placing the system on the market (Article 22), carry out conformity assessment, keep technical documentation, and register in the EU database.
- GPAI model provider? Article 54 requires an authorised representative established in the Union for providers of general-purpose AI models established in third countries — plus the documentation, copyright-policy and training-data-summary duties of Article 53.
- Only your EU customers deploy? Then the deployer duties (Article 26, oversight, logging, registration where applicable) sit with them — but they cannot discharge them without inputs from you: system documentation, intended purpose, classification reasoning and incident-reporting channels. Buyers treat your ability to supply these as a vendor requirement, contract or no contract.
- Substantially modify someone else’s system? Article 25 makes you the new provider. Fine-tuning and rebranding a US or EU model under your own name moves the full provider burden onto you.
What stays out: national security and defence activities fall to member-state discretion (Article 2(3)), and scientific research and development that precedes placing on the market is excluded. Neither exclusion rescues a commercial SaaS.
What EU buyers actually check
Three questions appear in every questionnaire: Who is your EU authorised representative? What is your classification for this system, and on what reasoning? How do you report serious incidents under Article 73? A US vendor that answers “not applicable — we are US-based” fails the first check instantly, because the reviewer knows the territorial rule.
The practical posture that passes: name your authorised representative, document the classification (even when the answer is minimal risk — the reasoning is what gets audited), and define the incident clock. That is a day of work, not a quarter.
Setting up without an EU entity
You do not need a subsidiary to comply. Three routes, in increasing order of commitment:
- Appoint an authorised representative — a EU-established person or company who holds your technical documentation and answers authorities on your behalf (Articles 22 and 54). Common, cheap, and expected.
- Contract an EU-based compliance owner — someone who signs off classification decisions and holds the evidence file. Works when your risk tier is limited or minimal.
- Incorporate in the EU — justified only at scale, or when you are a high-risk provider wanting first-line control of conformity assessment.
Whatever you choose, keep the evidence where an EU authority can reach it: versioned documents, dated classification records, and a named contact. The Act enforces through the member state where the output is used — being reachable there is the whole game.
How to check your own position in four minutes
The free exposure scan asks twelve questions about your AI usage and governance and returns an exposure score, your ranked exposures with the exact article that fired, and your first three moves — mapped to the dates that apply. Confirm your email and the full report unlocks.
Frequently asked questions
Does the EU AI Act apply to a US company with no EU presence?
Yes, if any of the Article 2 hooks apply — most commonly Article 2(1)(c): your system’s output is used in the Union. Establishment location is irrelevant; where the output lands is what counts. Expect to appoint an authorised representative established in the EU (Article 22 for high-risk systems, Article 54 for general-purpose AI models).
What is an authorised representative under the AI Act?
An EU-established natural or legal person formally mandated by a third-country provider to hold the technical documentation, provide it to authorities, cooperate with market surveillance, and complete registration steps on the provider’s behalf. They do not take over your product duties — they make you reachable to EU authorities.
Do we need an EU entity to comply with the AI Act?
No. Authorised-representative appointment plus an internal EU-facing compliance owner covers most configurations. Incorporation in the EU makes sense at scale or when you are a high-risk provider who wants direct control of the conformity process.
How does AI Act scope differ from GDPR for US firms?
The structure is deliberately parallel: scope follows users and output into the Union rather than establishment. The difference is in the duties — GDPR attaches to personal data processing, the AI Act attaches to placing and using AI systems, and adds product-style obligations (conformity assessment, technical documentation, registration) that have no GDPR equivalent.
What happens if a US company ignores the AI Act?
Enforcement runs through the member state where the system’s output is used, with penalties up to €15M or 3% of worldwide turnover for most obligation breaches (Article 99). In practice the first consequence is commercial: EU buyers cannot complete their own deployer-side documentation without your inputs, so unanswered questionnaires become lost deals long before an authority writes to you.
Put this guide to work on your own estate.
Run the free exposure scan to see which of these articles actually apply to your company — 12 questions, four minutes, no card. Confirm your email and the full report unlocks.
- Free forever plan
- No credit card
- EU-hosted · GDPR-native
Build a live AI register in minutes — upgrade only when you need the audit pack.
Keep reading
EU AI Act for SMEs and Start-Ups: What Applies, What's Exempt, and the Fine Rule That Helps You
9 min
ExplainersThe Digital Omnibus Delay: Your December 2027 High-Risk AI Deadline Is Already Ticking
12 min
ExplainersThe EU AI Act for Banks: Credit Scoring, Fraud Detection and Annex III Point 5
11 min