Malvertising Evolves to Weaponized Infrastructure, Bypassing Traditional Detection
Malvertising campaigns are increasingly sophisticated, shifting from deceptive ad content to complex, multi-stage redirect chains and cloaking techniques that hide malicious activity.
Malvertising campaigns are becoming significantly harder to detect by simply examining the ad creative itself. A growing proportion of malicious advertising activity now relies on post-click behaviors, employing intricate redirect chains, disposable domains, cloaking systems, and conditional delivery mechanisms. Attackers are increasingly hiding malicious components deep within the delivery chain, activating them based on specific user conditions rather than relying on overtly deceptive creatives.
Analysis of ad moderation data from PropellerAds, covering the first half of 2026, reveals a stark shift in the nature of rejected campaigns. While overall rejections decreased by 42%, malware and antivirus-flagged threats saw a 13% increase in absolute terms. More significantly, these technical threats grew from 23.3% to 45.9% of all rejections, becoming the dominant category. This rise occurred as simpler content violations, such as adult-content issues, were filtered out earlier in the moderation process.
This trend is corroborated by independent industry data. GeoEdge reported a rise in redirect-based attacks from 48% to 66% of malicious activity between Q1 and Q2 of 2025, while overall malicious ad activity doubled across major markets. Google's threat telemetry for the first half of 2026 further underscores this shift, indicating that malvertising accounted for nearly 30% of its threat detections, highlighting its pervasive overlap with the broader cybersecurity landscape.
Cloaking techniques remain a persistent tactic, accounting for a high percentage of advertiser suspensions. This indicates that hiding a campaign's true destination or behavior is a core operational strategy, not merely an occasional evasion method. The malicious asset is increasingly the entire delivery chain, rather than a single deceptive page.
The traditional ad review process, which focuses on visible elements like creatives and landing pages, is becoming less effective. Malicious behavior is often distributed across multiple technical components. A campaign might start with a seemingly compliant ad and a clean landing page, but then route users through tracking services, intermediate domains, conditional redirects, or additional pages before reaching the final malicious payload. Each step may appear harmless in isolation, with the risk only becoming apparent when the full path is reconstructed.
This distributed architecture offers significant advantages to malicious operators. It decouples the initial ad from the final payload, allowing individual domains to be replaced without dismantling the entire operation. Furthermore, traffic can be segmented by geography, device, or other signals, enabling different outcomes for different visitors. A recent malvertising campaign investigated by Confiant, active since late 2024, impersonated major brands and used visitor fingerprinting to deliver tailored experiences, showing that the campaign is effectively a delivery system rather than a single malicious object.
The economics of traffic also influence the form malicious advertising takes. In high-payout markets, higher CPC and CPA values can support greater investment in evasion infrastructure, such as longer redirect chains and more sophisticated cloaking. This economic logic favors persistence—surviving scrutiny to maximize value from each conversion. Conversely, in lower-cost, high-volume markets, cheaper, easily replicable, and discardable infrastructure becomes viable, with profitability driven by the sheer volume of attempts rather than heavily engineered campaigns.
Both models increasingly rely on complex infrastructure rather than a single deceptive page. Defenders must therefore focus on signals like destination switching, cloaking, and infrastructure reuse. The multi-hop redirect chains create a significant visibility problem, separating the point where a user interacts with an ad from the ultimate malicious destination, making traditional detection methods increasingly obsolete.