AIINEWS

AIINEWS

Global intelligence on artificial intelligence and computing

New security flaws allow data extraction from AMD virtual machines

AIINEWS Editorial Board•security
14 November, 2026•2 min read
New security flaws allow data extraction from AMD virtual machines

Executive Briefing

“Researchers have uncovered two new attack methods named Mior and Skyfall, allowing...”

Key Takeaways

  • 1.The Mior method developed by researchers from ETH Zurich allows intervention during virtual machine startup, enabling data reading and modification of memory as well as forging software attestation reports.
  • 2.The Skyfall method conducted jointly by experts from ETH Zurich and Google exploits transient caching behavior to recover register values and disk encryption keys.
  • 3.AMD has acknowledged these vulnerabilities and published a schedule to provide firmware and microcode patches for multiple EPYC processor lines in the coming months.

Researchers have uncovered two new attack methods named Mior and Skyfall, allowing malicious host servers to bypass memory encryption on AMD EPYC virtual machines protected by SEV-SNP. These vulnerabilities exploit the boot sequence or instruction timing to read sensitive data inside the virtual machines.

Key Developments and Technical Context

The Mior method developed by researchers from ETH Zurich allows intervention during virtual machine startup, enabling data reading and modification of memory as well as forging software attestation reports.

The Skyfall method conducted jointly by experts from ETH Zurich and Google exploits transient caching behavior to recover register values and disk encryption keys.

AMD has acknowledged these vulnerabilities and published a schedule to provide firmware and microcode patches for multiple EPYC processor lines in the coming months.

Strategic Analysis and Industry Implications

These attacks highlight the complexity of securing cloud environments for confidential computing technologies, especially when host operators control foundational firmware layers.

The reliance on cloud providers and hardware manufacturers to apply BIOS patches creates a significant challenge for customers renting confidential virtual machines.

Outlook and Key Factors to Watch

There is no confirmation from security advisories as to whether these vulnerabilities have been exploited in real-world attacks.

The specific effectiveness and methods for verifying trusted computing base values for virtual machine attestation after updates have not been detailed.

Content and editorial synthesis copyright by AIINEWS Editorial Board
Share this intelligence report:
Expanded Perspectives & Intelligence

Related Dispatches & Deep Dives

Curated 8 specialized reports
CrowdStrike to shell out nearly $500M for 2 AI security startups as the market for securing agents takes off

CrowdStrike to shell out nearly $500M for 2 AI security startups as the market for securing agents takes off

Cybersecurity giant CrowdStrike is buying two AI security companies that expand its products to more broadly monitor, test, and cover AI models and agents, as companies across industries start rolling out AI agents and models wholesale. CrowdStrike said it will buy Fiddler AI, which makes observability software for monitoring AI agents inside enterprises, and Dreadnode, which helps companies automate security testing specifically aimed at detecting flaws in AI models, apps and agents before they are set live. The company is paying around $400 million for Fiddler AI, and about $95 million for Dreadnode, according to The Wall Street Journal.

2 min read5,392
OpenAI launches GPT-6 Cyber, an Astra model built to keep vulnerability research in scope

OpenAI launches GPT-6 Cyber, an Astra model built to keep vulnerability research in scope

### OpenAI launches GPT-6 Cyber, an Astra model built to keep vulnerability research in scope OpenAI Group PBC today released GPT-6 Cyber, a version of its GPT-6 Astra model trained specifically for authorized vulnerability research. The new model goes after a problem OpenAI’s own security work exposed this summer. In July, agents running in the company’s test environments broke into Hugging Face Inc.

4 min read9,715

Explore AIINEWS Intelligence Vault

Over 2,000 in-depth dispatches structured across 18 specialized tech & AI categories.

Browse all 18 categories
New security flaws allow data extraction from AMD virtual machines | AIINEWS