On June 3, 2026, the European Commission published the Cloud and AI Development Act (CADA), the centerpiece of what Brussels is calling its European Technological Sovereignty Package. It introduces a mandatory four-tier sovereignty framework for cloud and AI services used by public bodies. The highest tier, covering defense and national security, would effectively exclude any cloud provider subject to foreign law.
The quote that cut through the press release noise came from Executive Vice-President Henna Virkkunen: the EU intends to ensure that critical cloud providers do not have a "kill switch." She said plainly that U.S. companies would have difficulty reaching the highest sovereignty tiers because of the U.S. CLOUD Act (18 U.S.C. § 2713), which allows the U.S. government to compel American companies to hand over data stored anywhere in the world.
Europe named the problem. The question is whether American regulated organizations - law firms, tribal governments, defense contractors, healthcare systems - are paying attention.
What CADA Does
The Cloud and AI Development Act defines cloud and AI sovereignty in four assurance levels:
Tier 1: Infrastructure physically located within the EU. Minimum baseline. All public bodies must meet this.
Tier 2: Demonstrated independence from third-country operational interference. Providers document what foreign law can compel them to disclose.
Tier 3: EU ownership, control, and personnel criteria. The provider's corporate chain must sit within EU jurisdiction.
Tier 4: Full supply-chain transparency and control with no third-country interference. Hardware, software, personnel, and legal jurisdiction must all be EU-originating. Reserved for defense, law enforcement, and border management - roughly 1% of public procurement by volume, but 100% of the most sensitive workloads.
The Commission estimates U.S. cloud companies currently control more than 70% of the EU cloud market. The EU currently spends €264 billion per year mostly on U.S. proprietary IT products and services. CADA is a structural intervention to change that dependency.
Why This Matters to American Regulated Organizations
You are not in the EU. CADA does not apply to you directly. So why does this matter?
Because CADA is the first formal legal definition of the structural problem that American regulated industries have been quietly tolerating for years: the extraterritoriality problem.
The U.S. CLOUD Act (18 U.S.C. § 2713) requires American cloud providers to disclose customer data when compelled by U.S. law, regardless of where that data is stored. This means that when a tribal government uploads operational data to AWS, a law firm runs a contract review through Microsoft Azure AI, or a defense contractor feeds CUI into a cloud LLM, that data sits under a legal regime that their own federal government can access through a provider-level order - without the customer's knowledge or consent.
Europe drew a tier around that. It is called Tier 4. It says: if your provider is subject to CLOUD Act jurisdiction, you cannot use it for your most sensitive work.
American regulators have not drawn that line. But the underlying legal exposure is identical.
The Regulatory Reality Your Cloud Provider Won't Discuss
Tribal Nations: OCAP and Data Sovereignty
The First Nations Principles of OCAP (Ownership, Control, Access, and Possession) establish that a tribe owns its data collectively, controls how it is used, must be able to access it at will, and must physically possess it. The CLOUD Act directly violates the Possession principle. A tribal government running administrative AI through any U.S.-headquartered cloud provider has, by operation of law, surrendered physical possession of that data to a legal regime it does not control.
For a deeper analysis of how OCAP and the CLOUD Act interact, see our post on OCAP Principles and the CLOUD Act: A Compliance Guide for Tribal Governments. For the structural argument on why cloud AI and tribal data sovereignty are incompatible by design, see Tribal Data Sovereignty and the Cloud AI Problem.
Law Firms: ABA Model Rule 1.6
ABA Model Rule 1.6 requires a lawyer to make reasonable efforts to prevent the inadvertent or unauthorized disclosure of client information. The word "reasonable" carries weight. As of 2026, with sovereignty-compliant local AI infrastructure commercially available, running client data through a third-party cloud AI service is an increasingly difficult position to call reasonable. The EU just published a four-tier framework explaining why.
For the full structural analysis of how privilege breaks in cloud AI workflows, see our post on Attorney-Client Privilege and Cloud AI: The Structural Problem Law Firms Can't Negotiate Away. For law firm AI infrastructure options, see our solutions page.
Defense Contractors: ITAR, DFARS, and CMMC
International Traffic in Arms Regulations (22 C.F.R. Parts 120-130) restrict the export of controlled technical data to foreign nationals. Cloud providers are subject to compelled disclosure under the CLOUD Act. If that compelled disclosure involves a foreign national employee or a foreign-jurisdiction data center, the contractor faces an unauthorized export. CMMC 2.0 requires protection of Controlled Unclassified Information (CUI) under NIST SP 800-171. Air-gapped local infrastructure is the only technical control set that fully satisfies the isolation requirement.
See our ITAR and DFARS AI Self-Assessment for a structured evaluation, and our defense contractor AI solutions page for configuration options.
Healthcare: HIPAA Technical Safeguards
45 C.F.R. § 164.312 requires covered entities to implement technical security measures that prevent unauthorized access to ePHI transmitted over electronic communication networks. A Business Associate Agreement (BAA) with a cloud AI provider transfers contractual liability but does not transfer physical control of the data. The CLOUD Act can compel disclosure of ePHI from a cloud provider regardless of the BAA.
For the full HIPAA technical safeguards analysis, see our HIPAA Technical Safeguards for Local AI Deployment post, and our medical practice AI infrastructure page.
What the EU's Framework Reveals About Your Current Setup
CADA's four-tier model is useful precisely because it forces an honest answer to a question most American organizations avoid: at what tier is your current AI deployment?
If you are using any major cloud AI service - OpenAI's API, Microsoft Azure AI, AWS Bedrock, Google Vertex AI - you are at Tier 1 at best under the CADA framework. Infrastructure may be domestic. But your provider is subject to U.S. law, is run by a publicly traded company with shareholders who are not your clients, and operates under terms of service that can change without your approval.
For regulated industries handling data that has legal protection attached to it - PHI, CUI, client confidences, tribal proprietary data - Tier 1 is not a compliant posture. It is a liability waiting to materialize.
The Technical Answer: Local Inference Infrastructure
Island Mountain builds pre-installed, burn-tested, air-gapped inference servers using NVIDIA H100 and H200 GPUs. The hardware ships to your building. Your data never leaves your premises. There is no vendor agreement governing data access because there is no data transmission. The CLOUD Act cannot compel production of data that a cloud provider does not have.
Summit Base ($75,000-$85,000): Entry-level production inference server. Handles most regulated-industry workloads including document review, summarization, structured data analysis, and contract analysis. Ships with a 72-hour burn-in test completed. Pre-installed model stack includes DeepSeek V4-Flash (quantized), Llama 4 Scout, and a buyer-selected third model.
Summit Ridge ($150,000-$160,000, build-to-order): Expanded configuration for organizations running heavier concurrent workloads or multiple department deployments. Custom GPU, CPU, and RAM options matched to your use case. Contact us to spec the right configuration.
Summit Pinnacle ($350,000-$400,000, (pre-order now): H200-based configuration for organizations requiring maximum inference throughput at scale. 282GB VRAM total. DeepSeek V4-Flash at full native quality and 1M context window.
Each unit ships with network configuration and an initial setup session. No token fees. No per-query billing. No cloud connectivity required. For a full five-year cost comparison against cloud AI spending, see Cloud AI vs. Local Hardware: The Honest 5-Year TCO.
What You Do Not Get
Local inference hardware is not a compliance certification. Buying a Summit Base does not make you HIPAA-compliant, CMMC-certified, or ITAR-registered. Those certifications require policies, training, audits, and documented controls across your organization's full operations.
What local hardware provides is a specific technical control: your data does not leave your premises. That control eliminates one category of risk that cloud AI introduces - third-party data custody - while leaving the rest of your compliance obligations intact.
If your legal or compliance team tells you that a BAA with OpenAI solves your HIPAA exposure, they may be right about the contractual liability. They are wrong about the technical control. CADA's framework is useful here: the EU explicitly states that data transfer frameworks do not remove sovereignty concerns because sovereignty goes beyond data transfers and relates to operational autonomy.
Read that sentence again. The EU Commission just told the world that contractual frameworks do not solve infrastructure sovereignty problems.
The Convergence Happening Right Now
This week, two things happened simultaneously that anyone in regulated-industry AI should track.
June 2: GIS Reports published a detailed analysis noting that governments worldwide are treating AI as a critical technology and imposing state control over its deployment. The Anthropic case, in which the Pentagon attempted to designate a U.S. AI company a national security supply chain risk for declining to support autonomous lethal weapons, illustrates that frontier cloud AI models are now geopolitical assets. What you query, and through which provider, is no longer a purely technical decision.
June 3: The EU published CADA, creating formal legal definitions for what "sovereignty" means in an AI context. The framework is now codified.
The direction of travel is clear. Regulated organizations that keep deferring the decision to bring AI in-house are deferring into an increasingly crowded regulatory corner.