FortiBleed: Leaked admin credentials expose nearly 74,000 Fortinet firewalls worldwide

A massive credential dataset linked to Fortinet firewalls highlights the risks of credential reuse, weak password hygiene and exposed management interfaces

A recently uncovered dataset known as “FortiBleed” contains user and administrator credentials for 73,932 Fortinet firewall (FortiGate) instances across 194 countries, according to multiple security researchers. Analysts estimate that this could represent a significant share of internet-facing Fortinet systems globally - with some suggesting it affects roughly half of the FortiGate devices discoverable via Shodan.

Many of these credentials are believed to remain valid. For affected organisations, this raises the possibility that attackers may already have gained access to corporate networks using legitimate login details. The data spans organisations across numerous sectors, including major enterprises such as Mercedes-Benz, Siemens, Oracle and Samsung, as well as public sector entities.

The exposure was identified by security researcher Volodymyr “Bob” Diachenko, who discovered an openly accessible server linked to an attacker group. Evidence points to an ongoing, highly automated campaign that does not rely on a new vulnerability, but instead exploits compromised credentials, weak password hygiene, and poorly secured management and VPN interfaces.

Attack method and root causes

According to analysis from SOCRadar, Hudson Rock and others, the actors behind FortiBleed are using automated large-scale scanning, credential stuffing and infrastructure designed to crack password hashes. They begin by identifying internet-exposed Fortinet firewalls with accessible management or SSL VPN interfaces, then systematically test known login credentials - including those from earlier Fortinet breaches, public leak databases and infostealer malware.

Diachenko reported around 1.16 billion credential attempts across more than 320,000 FortiGate targets. Attackers are also believed to have intercepted authentication hashes from active SSL VPN sessions and cracked them offline using GPU clusters and tools such as Hashtopolis.

This approach allows even long, complex passwords to be compromised if they are stored using outdated or fast hash formats. Fortinet has since migrated administrator passwords in newer FortiOS releases to the more resilient PBKDF2 algorithm.

However, documentation indicates that after upgrades, existing passwords may remain stored in the older SHA-256 format until users log in again or reset their credentials. As a result, even updated systems may still contain vulnerable legacy hashes. Systems where passwords have not been changed for extended periods, or where dormant admin accounts persist, are particularly at risk.

The campaign reflects a straightforward but highly effective strategy: credential reuse rather than exploitation of a new zero-day vulnerability. While no new Fortinet flaw has been confirmed, researchers believe the dataset draws on prior breaches, infostealer logs and systematic credential harvesting. FortiBleed therefore represents a large-scale exploitation of weak credential hygiene rather than a traditional software exploit.

Fortinet’s response and official guidance

Fortinet has stated that it is aware of reports describing a credential-harvesting campaign targeting its firewalls and VPN gateways. The company indicated that the dataset likely aggregates older incidents alongside credentials obtained via brute-force attacks, and that it does not relate to any newly disclosed vulnerability.

Some independent researchers dispute aspects of that assessment. Kevin Beaumont has verified selected administrator logins and noted that many affected systems remain online and distinct from earlier leaks.

Security firms including SOCRadar and Hudson Rock describe FortiBleed as a large-scale, ongoing credential harvesting operation targeting FortiGate and SSL VPN environments. SOCRadar has released a checker to help organisations determine whether their systems are affected.

Fortinet products have repeatedly featured in official vulnerability advisories and international KEV listings in recent years.

The mitigation priorities are clear:

Risks and context

Unlike previous Fortinet incidents - such as the 2021 leak of around 500,000 FortiGate VPN credentials or the “Belsen” dataset in early 2025 - FortiBleed does not centre on a single known vulnerability.

Instead, it demonstrates how poor password management, patching and system exposure can be exploited at industrial scale. The risk is particularly acute at the network perimeter: if attackers gain control of a FortiGate firewall using valid credentials, it can serve as a launch point for further compromise. In some cases, compromised devices have reportedly been used as “listening posts” to capture additional credentials from passing traffic.

This creates a self-reinforcing attack chain, where one compromised credential leads to broader access - including Active Directory, RADIUS or VPN accounts - and facilitates lateral movement. By the time credentials are reset, attackers may already have established persistence, created new accounts or exfiltrated data. Researchers have reported full network compromises in several countries, including the suspected theft of sensitive documents from a defence contractor in Turkey.

For organisations - particularly in regulated or critical sectors - the incident underscores the need to secure perimeter systems rigorously and implement zero trust principles in practice, not just strategy.

Multi-factor authentication should be standard across all externally accessible management and VPN endpoints. Organisations also need strict access controls, removal of unnecessary internet-facing interfaces, frequent password rotation, rapid patching cycles, network segmentation and continuous monitoring of security logs.

Even fully patched systems with strong passwords remain vulnerable if management interfaces are exposed or compromised credentials are not replaced quickly. FortiBleed ultimately reinforces a longstanding lesson: credentials remain one of the weakest links in enterprise security - and robust identity and access controls are critical to preventing large-scale compromise.

This article was first published on Computing Deutschland