Skip to main content
Mallory
← Back to News
post-quantum-cryptographyai-platform-securityembedded-device-vulnerability

Quantum-Era Security Discourse and Side-Channel Risks for PQC-Enabled AI Inference

Emerging Technology Risk · 2 sources · Updated 5mo ago · First seen Feb 6, 2026

Quantum-Era Security Discourse and Side-Channel Risks for PQC-Enabled AI Inference

A Vulnerable U episode previewed content filmed at Palo Alto Networks focused on how quantum computing is expected to impact security planning over the next decade, arguing that while the underlying physics is complex, many preparedness actions resemble familiar infosec transitions (e.g., inventorying cryptography dependencies and planning migrations). It also highlighted how public debate around encryption can be distorted by headlines—citing renewed discussion tied to a WhatsApp-related lawsuit and pointing to cryptographer Matthew Green’s technical breakdown as a corrective to overbroad claims that “end-to-end encryption is broken.”

Separately, Gopher Security warned that adopting post-quantum cryptography (PQC) does not address implementation-layer leakage, emphasizing that side-channel attacks (including remotely observable timing/power effects such as Hertzbleed) can undermine “quantum-safe” designs. The post described how electromagnetic (EM) emissions from AI hardware can leak sensitive information during inference, citing research such as BarraCUDA (weight extraction from NVIDIA Jetson devices via EM measurements) and academic work indicating side-channels can break embedded AI “black-box” assumptions by enabling extraction of logits/weights or gradient estimation—creating model theft and evasion risks even when cryptographic algorithms are strong.

New surface, new adversariesTrack how attackers are adapting to this technology.
Track this risk
Event timeline

How this story unfolded

Events from the most recent confirmed update back to the earliest known activity.

10 events · Jun 14, 2022 – Feb 6, 2026
  1. Feb 6, 202610mo ago

    Court filing says iPhone Lockdown Mode blocked FBI access

    A court filing indicated that Apple iPhone Lockdown Mode prevented the FBI from accessing a reporter’s iPhone, illustrating the feature’s real-world defensive impact.

  2. Feb 6, 202610mo ago

    Unit 42 details ShadowGuard and Diaoyu espionage campaign

    Palo Alto Networks Unit 42 described a broad, geopolitically aligned campaign using the ShadowGuard eBPF kernel rootkit together with a custom loader called Diaoyu.

  3. Feb 6, 202610mo ago

    China-linked espionage campaign exploits WinRAR vulnerability

    Reporting described China-linked espionage activity leveraging a WinRAR vulnerability as part of ongoing operations.

  4. Feb 6, 202610mo ago

    APT28 rapidly exploits newly patched Microsoft Office flaw

    Russia-linked APT28 was reported to have quickly begun exploiting a newly patched Microsoft Office vulnerability after a fix became available.

  5. Feb 6, 202610mo ago

    China-linked actors reportedly compromise Notepad++ update infrastructure

    A reported supply-chain compromise affected Notepad++ update infrastructure, with the activity attributed to China-linked attackers.

  6. Feb 6, 202610mo ago

    Attackers actively exploit React Native Metro RCE bug

    Reports indicated active exploitation of a React Native Metro vulnerability that allows remote command execution against exposed development servers.

  7. Feb 6, 202610mo ago

    CISA warns active exploitation of GitLab SSRF flaw

    CISA warned that an older GitLab server-side request forgery vulnerability was being actively exploited in the wild, elevating the urgency for organizations to patch exposed systems.

  8. Jan 1, 20252y ago

    One-trace horizontal attack reported against mkm4 Kyber on Cortex-M4

    A reported side-channel attack demonstrated secret extraction from a Cortex-M4 implementation of Kyber (mkm4) using a single trace, highlighting implementation risks in post-quantum cryptography.

  9. Jan 1, 20252y ago

    BarraCUDA research shows EM extraction of AI model weights

    Researchers reported that electromagnetic emissions from NVIDIA Jetson devices during inference could be used to recover neural network weights, undermining the assumption that deployed models are only accessible as black boxes.

  10. Jun 14, 20224y ago

    Hertzbleed demonstrates remotely observable power-related leakage

    The Hertzbleed research showed that power-related side-channel leakage can be observed remotely through runtime effects, challenging the assumption that side-channel attacks require physical access.

Linked entities

Related entities

Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.

Open in app →

Affected products

1 linked
Whatsapp

Organizations

1 linked
NVIDIA Corporation
Do you run this?Follow how adversaries are adapting to this technology, and where it touches your stack today.
Track This Risk
Mallory platform

See the full picture, correlated to your attack surface.

This page covers what’s public. Mallory adds the parts that aren’t: which of your assets are affected, which threat actors are using it right now, and what to deploy next.

Exposure mappingMap indicators from this story to your assets and identify affected systems in minutes.
Threat actor evidenceEvery observed campaign, victim, and pivot linked to actors named in this story.
Associated malwareMalware, exploits, and IOCs connected to the activity described here.
Detection signaturesYARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Scheduled alertsGet matching new stories delivered to your team as they break, not the next morning.
AI threadsAsk questions about this story and take action on the answers.