Europe built sovereign clouds to escape US control. Forgot about the processors

TL;DR

Prime Big Deal Days · Oct 6–7Offer from Amazon

Get the latest gadgets delivered free — and shop member deals

• Fast, free delivery on millions of items
• Access to Prime Big Deal Days deals on October 6–7
• Prime Video, Amazon Music and more included

Start your free Prime trial →Free trial for eligible customers · Cancel anytime

As an affiliate, we earn on qualifying purchases.Hide

Europe has invested over €2 billion in sovereign cloud initiatives to reduce US legal influence. However, most data centers still rely on Intel and AMD processors, which contain management engines operating below the OS, posing security risks. The situation raises questions about true digital sovereignty and hardware security.

European efforts to establish sovereign cloud infrastructure are challenged by the continued reliance on Intel and AMD processors, which contain embedded management engines operating at a privilege level inaccessible to the host system, raising security and sovereignty concerns.

Europe has allocated more than €2 billion through the EU’s IPCEI-CIS program to develop infrastructure that ensures data sovereignty and immunity from extraterritorial laws. France, among other nations, has adopted frameworks like SecNumCloud, which set high standards for security and legal immunity. Despite these efforts, most data centers and qualified cloud operators still depend heavily on US-made processors from Intel and AMD.

Inside these processors are management engines—the Intel Management Engine (ME) and AMD Platform Security Processor (PSP)—which operate independently at a privilege level below the operating system, known as Ring -3. These microcontrollers have their own memory, network stack, and clock, making them capable of running covert operations without the host OS’s knowledge. Researchers like John Goodacre have described these as “a computer inside your computer,” capable of maintaining persistent access even when the device appears powered off.

Security experts point out that these management engines can be exploited for covert channels, remote management, and even backdoors. Historical cases, such as the 2017 documented use of Intel’s Serial-over-LAN (SOL) by nation-state actors, demonstrate how these features can be exploited for espionage and data exfiltration. Recent research shows vulnerabilities in AMD’s SEV-SNP technology, with exploits achieving a 100% success rate in controlled tests, further highlighting risks.

Why It Matters

This situation is significant because it questions the effectiveness of Europe’s sovereignty initiatives. While legal and infrastructural measures aim to insulate data from US jurisdiction, the underlying hardware architecture—specifically the embedded management engines—remains a potential security loophole. If these engines can be exploited or tampered with, the entire premise of digital sovereignty could be compromised.

Moreover, the presence of covert channels and backdoors in widely used processors means that even compliant, certified cloud providers might be vulnerable to espionage, sabotage, or unauthorized data access. This undermines trust in European cloud infrastructure and complicates efforts to achieve truly independent and secure digital ecosystems.

Background

European countries have been increasingly investing in sovereign cloud projects since the EU launched initiatives like IPCEI-CIS, aiming to reduce dependence on US technology and legal jurisdiction. France’s SecNumCloud framework sets strict security standards, but hardware reliance on US silicon persists. The embedded management engines in Intel and AMD processors have been known for years, with documented vulnerabilities and potential for covert operations, including espionage by state actors.

Historically, nation-states like the US have exploited these features for intelligence gathering. Recent research, such as the Fabricked attack against AMD’s SEV-SNP, demonstrates that vulnerabilities in these processors are not just theoretical but practically exploitable, raising concerns about the security of European cloud infrastructure built on such hardware.

“It’s a computer inside your computer. The Management Engine has its own memory, its own clock, and its own network stack, operating below the host system’s visibility.”

— John Goodacre

“Yes, it can probably be used as a backdoor, like many other firmwares. The real question is whether operational controls can make it unreachable in practice.”

— Professor Aurélien Francillon

What Remains Unclear

It remains unclear how widespread the exploitation of these management engines currently is and whether European cloud providers are actively mitigating these hardware risks. The full extent of vulnerabilities in AMD’s SEV-SNP and other confidential computing technologies is still being evaluated, and hardware manufacturers have not universally addressed these concerns.

What’s Next

Next steps include further research into hardware vulnerabilities, increased scrutiny of management engine security, and potential development of processors without such embedded management features. European regulators and cloud providers may also seek to implement hardware-level security measures or shift towards alternative architectures that exclude US-based silicon.

Key Questions

Why are management engines a security concern?

Management engines operate below the host operating system, with their own memory and network capabilities, making them difficult to monitor or control. They can be exploited for covert data exfiltration, remote management, or backdoors, posing risks to security and sovereignty.

Can Europe build fully sovereign cloud infrastructure without US processors?

While Europe aims to do so, current technological limitations mean most infrastructure still relies on US processors with embedded management engines. Developing or sourcing processors without such features is a key challenge for achieving true sovereignty.

What are the risks of exploiting these hardware vulnerabilities?

Exploiting management engines can enable covert espionage, remote control, data theft, or sabotage of cloud infrastructure, undermining data security and privacy, and potentially violating sovereignty agreements.

Are there any hardware alternatives to US-based processors?

Some efforts are underway to develop processors free of management engines or built with hardware-based security features, but such solutions are not yet widespread or commercially mature for large-scale deployment.