New research into the Lazarus cyberespionage group reveals that attackers are already using NIST-standardised post-quantum cryptography to protect malicious communications. The discovery demonstrates that quantum-safe security is entering operational use, although its adoption by cybercriminals raises a broader question about the relationship between strong encryption, cybersecurity and digital trust.
The transition to post-quantum cryptography is generally presented as a defensive undertaking. Governments, technology providers and enterprises are preparing to replace vulnerable cryptographic systems before sufficiently powerful quantum computers can compromise them. But new research from Kaspersky illustrates that the same technologies are also finding their way into the hands of sophisticated cyberattackers.
On 21 October, Kaspersky’s Global Research and Analysis Team (GReAT) published details of Operation FlashHole, a cyberespionage campaign attributed to the North Korean-linked Lazarus group. Among its findings was the use of ML-KEM-1024, one of the parameter sets defined in the post-quantum cryptographic standard published by the US National Institute of Standards and Technology (NIST).
The mechanism was incorporated into malware communications, allowing the attackers to establish shared cryptographic secrets with their command-and-control infrastructure before receiving and decrypting additional malicious software.
For an industry focused on preparing legitimate systems for the quantum era, the discovery provides an important reminder: advances in cryptographic security are available to everyone, including those operating on the other side of the security equation.
A new capability in an established cyberespionage operation
According to Kaspersky, Operation FlashHole has been active since at least July 2025 and has affected at least three organisations in South Korea, operating in construction, software development and communications equipment manufacturing.
The campaign combined established attack techniques with newly developed malware and more sophisticated methods of concealment. Researchers identified spear-phishing emails directing targets towards fraudulent professional websites, including fabricated résumés and portfolio pages. One site contained an AI-generated photograph of a supposed job applicant, while researchers believe AI may also have been used to customise fraudulent websites for individual targets.
The attackers exploited the trust associated with apparently legitimate professional material and abused components of CrossWeb EX, security and web-authentication software used in South Korea, as part of the infection process.
Once access had been established, the malware chain used a combination of established and previously undocumented tools, including TouchHub, ClientHub and an updated version of LPEClient. It was within LPEClient that researchers identified the post-quantum cryptographic mechanism.
The malware gathered information about compromised computers, including installed security software and Windows protections, before communicating with the attackers’ infrastructure and receiving further encrypted payloads. The attack chain also employed memory-based execution and techniques intended to interfere with security monitoring.
The most significant finding for the wider cryptographic industry concerns the incorporation of ML-KEM-1024 into this process. Kaspersky’s investigators say they can trace the group’s use of post-quantum cryptography back to at least September 2024, indicating that the capability predates the latest disclosure by more than two years.
Why ML-KEM matters
ML-KEM, the Module-Lattice-Based Key-Encapsulation Mechanism, was standardised by NIST in August 2024 under FIPS 203. Derived from the CRYSTALS-Kyber algorithm, it enables two parties to establish a shared secret over an otherwise untrusted communications channel.
It is designed to resist attacks from both conventional computers and future cryptographically relevant quantum computers. Its three standardised parameter sets, ML-KEM-512, ML-KEM-768 and ML-KEM-1024, offer different levels of security and performance.
The technology is central to the wider migration towards quantum-safe communications. Its adoption by cloud providers, software vendors, semiconductor manufacturers and enterprise security platforms is part of a growing effort to reduce dependence on public-key algorithms that could eventually become vulnerable to quantum attacks.
In Operation FlashHole, however, ML-KEM-1024 was serving a different purpose. Rather than protecting sensitive business information, financial transactions or legitimate system communications, it helped establish cryptographic keys used to protect the delivery of malicious payloads.
This does not indicate a weakness in ML-KEM or suggest that the underlying standard has been compromised. Quite the opposite: attackers appear to recognise the value of strong cryptographic protection for their own infrastructure.
There is also no evidence in these findings that Lazarus possesses a quantum computer or has developed a quantum capability for breaking encryption. The group is using software implementing a publicly standardised post-quantum algorithm.
That distinction matters. The development concerns the adoption of quantum-resistant cryptography, not an increase in attackers’ quantum computing capabilities.
Strong encryption does not establish trust
The findings expose a distinction that can become obscured in discussions about quantum-safe security.
Cryptography protects information and, when combined with appropriate protocols and authentication, can help establish trusted communications. It cannot independently determine whether the parties using that protection are legitimate.
A malicious command-and-control server can benefit from strong encryption in much the same way as a legitimate enterprise service. Both have reasons to prevent outsiders from reading their communications or interfering with transmitted information.
For defenders, this creates an additional challenge. Encrypted traffic already complicates efforts to understand malicious activity, particularly when attackers combine it with legitimate cloud services, compromised software and memory-resident malware.
The introduction of post-quantum cryptography does not fundamentally create that problem, but it demonstrates that attackers are willing to adopt emerging standards when doing so supports their operations.
It also complicates the assumption that the cybersecurity industry’s migration towards quantum-safe cryptography will proceed entirely as a planned, defensive transformation.
Threat actors can adopt new technology without waiting for enterprise migration programmes, regulatory deadlines or industry-wide interoperability agreements.
What this means for the PQC transition
For organisations preparing their own migration, the immediate priority remains unchanged. Cryptographic inventories, dependency assessments, migration planning and the introduction of standardised quantum-resistant mechanisms are necessary steps towards protecting information over the longer term.
The Lazarus findings do not weaken that case. They do, however, reinforce the need to understand post-quantum migration as one part of a much broader security architecture.
Cryptographic resilience must operate alongside identity assurance, secure software execution, endpoint protection, behavioural monitoring and effective incident response. The strength of the encryption used within a system cannot compensate for compromised credentials, manipulated applications or malicious code running with legitimate permissions.
There is a further implication for software and security providers. As post-quantum mechanisms become increasingly common in legitimate network traffic, security tools will need to distinguish authorised activity from malicious behaviour without relying on the assumption that an unfamiliar cryptographic protocol is necessarily suspicious.
For those developing security products, cryptographic agility must therefore extend beyond supporting new algorithms. It must include the ability to manage, monitor and investigate the systems in which those algorithms are deployed.
The industry is entering a period in which quantum-resistant cryptography will increasingly become part of ordinary computing infrastructure. It is reasonable to expect attackers to make use of that infrastructure as well.
A different measure of quantum readiness
Operation FlashHole is unlikely to change the timetable for enterprise PQC migration, and the number of identified victims remains limited. But the significance of the research extends beyond the immediate campaign.
Much of the discussion surrounding quantum-safe security focuses on a future threat: the possibility that powerful quantum computers will eventually undermine widely used cryptographic protections.
The Lazarus case demonstrates something happening in the present. Post-quantum cryptography has already become sufficiently accessible and operationally useful to appear in sophisticated offensive cyber operations.
That makes the development relevant well beyond the organisations directly affected. For enterprises, quantum readiness cannot be measured solely by how many vulnerable algorithms have been replaced. It must also reflect the ability to maintain visibility, control and accountability as the technologies protecting digital systems evolve.
The security of a cryptographic mechanism and the trustworthiness of the organisation using it remain two different questions. As post-quantum cryptography moves into everyday deployment, recognising that distinction will become increasingly important.
Leave a Reply