How Does the SANSAR Malware Hide Mining on AI GPU Servers?

Article Highlights
Off On

The rapid expansion of distributed computing clusters has inadvertently created a massive attack surface that sophisticated threat actors are now exploiting with surgical precision. As the industry moves further into 2026, the SANSAR malware campaign stands out as a prime example of how traditional cryptojacking has evolved into a highly specialized form of infrastructure subversion targeting high-end GPU servers. Attributed to the group identified as REF3364, this toolkit does not simply run unauthorized code; it fundamentally compromises the integrity of the host environment to ensure a long-term, profitable presence. Unlike the noisy, resource-hogging miners of previous years, SANSAR is engineered for deep stealth, prioritizing the subversion of monitoring tools over sheer speed. This represents a significant shift in the threat landscape, where the value of a compromised system lies not just in its data, but in its massive, unmonitored computational power. Security teams are finding that standard detection methods are insufficient against such layered deception.

Exploiting Weaknesses in AI Infrastructure

Modern AI research often relies on frameworks like Ray to manage complex workloads across hundreds of individual nodes, yet these tools frequently lack the robust security defaults necessary for internet-facing deployments. The SANSAR campaign begins by identifying internet-accessible Ray job agents that have been left without proper authentication, allowing attackers to treat the cluster as their own personal supercomputer. Once they establish a foothold, the threat actors initiate a Python-based reverse shell that provides a persistent gateway into the infrastructure. This approach is particularly effective because it leverages the very tools intended for legitimate high-performance computing, making the initial intrusion look like a standard administrative task. By using these legitimate channels, the attackers can bypass perimeter security measures that are typically tuned to watch for external exploits or phishing attempts. This initial compromise is the first step in a process designed to turn expensive hardware into a silent revenue stream.

Building on this foundation of unauthorized access, the toolkit then begins the process of “living off the land” by utilizing internal Python environments and system utilities to avoid raising red flags. This strategy is highly effective because it minimizes the introduction of external, easily detectable binary files during the early stages of the attack. Moreover, by operating within the context of a distributed computing framework, the malware can spread across multiple nodes with minimal effort, effectively hijacking the entire cluster’s capacity. This lateral movement is facilitated by the inherent trust built into these frameworks, where nodes are often configured to communicate freely to share data and processing tasks. Consequently, an organization might see a perfectly normal distribution of tasks while, in reality, every cycle is being diverted to solve cryptocurrency hashes. This highlights a growing need for security policies that treat internal compute job submissions with the same level of scrutiny as external traffic.

Manipulating GPU Telemetry and Metrics

The true technical brilliance—and danger—of the SANSAR toolkit lies in its ability to manipulate the NVIDIA Management Library (NVML), which serves as the primary data source for hardware health. When an administrator runs commands like nvidia-smi to check on the status of their expensive H100 or A100 clusters, they expect to see accurate utilization percentages and power draws. However, the malware replaces the standard library with a malicious wrapper that intercepts these queries and returns completely fabricated results. Even if the GPUs are operating at maximum capacity to mine Monero or other coins, the reporting tool will display a convincing 0% utilization rate. This tactical gaslighting makes it nearly impossible for traditional monitoring dashboards to trigger alerts based on resource exhaustion. By controlling the flow of information from the hardware to the operating system, the attackers ensure that their presence remains undetected by even the most vigilant human operators who trust their telemetry.

To further bolster this illusion of an idle system, the malware authors have implemented a sophisticated system of randomized environmental telemetry. It is not enough to simply report no activity; a truly idle server would also show lower power consumption and cooler operating temperatures. SANSAR addresses this by injecting plausible, randomized values into the NVML wrapper, reporting power usage between 100 and 200 watts and temperatures that hover comfortably between 36 and 46 degrees Celsius. This level of detail is designed to fool sanity checks that might look for high heat or fan noise as indicators of hidden workloads. In contrast to simpler malware that only hides processes, SANSAR creates a comprehensive, fake reality for the system administrator. This evolution in malware design suggests that attackers are no longer content with just staying out of the process list; they are actively rewriting the hardware’s narrative to persist for months without ever being questioned by maintenance teams.

Strategic Defenses Against Specialized Rootkits

Maintaining long-term access requires more than just falsified metrics; it necessitates a deep-rooted presence within the Linux kernel itself. SANSAR achieves this through a rootkit compiled using Dynamic Kernel Module Support, which allows the malware to adapt to whatever kernel version is currently running on the host. This kernel-level integration provides the attacker with the power to intercept system calls, effectively making certain files, network sockets, and processes completely invisible to the user space. For instance, a standard directory listing will never reveal the presence of the mining software or the configuration files used to manage the operation. Furthermore, the toolkit utilizes a highly resilient command-and-control architecture that uses Ethereum smart contracts to update its server addresses. By anchoring its infrastructure in the blockchain and routing all traffic through the Tor network, the campaign ensures that its communication channels are nearly impossible to block through traditional IP-based blacklisting or DNS filtering. Protecting high-performance environments from such advanced threats requires a fundamental shift toward verifying the integrity of the monitoring tools themselves. It has become clear that simply trusting the output of a hardware library is a vulnerability that attackers can exploit with devastating efficiency. To counter this, security professionals are moving toward implementing eBPF-based monitoring solutions that can detect unauthorized kernel hooks and system call interceptions at the source. Furthermore, the industry is prioritizing the strict authentication of distributed computing frameworks, ensuring that job agents are never exposed to untrusted networks. Organizations are also adopting external power monitoring and independent hardware sensors that function outside the compromised operating system to detect discrepancies in energy usage. By treating hardware telemetry as a claim that must be verified rather than an absolute truth, defenders can establish a more resilient posture against invisible cryptojacking campaigns that seek to subvert the AI era.

Explore more

How Will Agentic AI Transform Financial Operations?

Rules-based systems provide the precision of a rhythm section, while agentic AI acts as a soloist navigating complex judgment calls. In the current landscape of 2026, the financial sector has long relied on automation to boost efficiency, yet traditional methods often fail to address the inherent friction between siloed tasks. While individual steps may move faster, the overall workflow frequently

How Can Agency Leaders Foster Meaningful Career Growth?

Providing a safety net of feedback while allowing employees to navigate unfamiliar problems is a key component of guided delegation and professional mentorship. In the current landscape of advertising and public relations, the distinction between a heavy workload and genuine professional advancement has become increasingly blurred. Agency leaders frequently fall into the trap of assuming that by simply piling more

Toss Launches Stablecoin Wallet to Mainstream Digital Payments

Most merchants currently lack the technical stacks required for on-chain acceptance, which include hosted checkouts, liquidity management, and strict adherence to anti-money laundering protocols. This fundamental technological barrier is being challenged by the recent rollout from Toss, the South Korean fintech super-app, which integrated a robust stablecoin wallet on November 16, 2026. This development marks a transition in which digital

How Buy Now Pay Later Affects Your Future Loan Applications

Closing an unused credit account today is more beneficial than closing it right before a loan application because the record persists for two years. The buy now, pay later account opened at a digital checkout this month can remain on a credit report when a home, car, or business loan is assessed years from now. Since June 2025, Australian law

Is Embedded Finance Evolving Beyond Hype into Utility?

The rapid transition from conceptual financial novelties to essential operational infrastructure has fundamentally altered how businesses interact with software platforms in the current economic landscape. According to the latest Merchant Insider Report, which surveyed over fifteen hundred businesses across the United States, the United Kingdom, and Australia, the presence of financial tools within a software ecosystem is no longer a

Fill out the form below