Zero-day exploits are no longer an elite-only capability
Zero-days have long carried a mystique: an undisclosed flaw, an expensive exploit and a state-backed operator. But FireEye’s analysis of observed exploitation in 2019 challenged the idea that only highly sophisticated advanced persistent threat (APT) groups use them. Dark Reading reported that FireEye saw more attackers using zero-day exploits in 2019 than in the preceding comparison year. The finding is about observed activity; it does not mean zero-days are easy to obtain or that every kind of attacker used them. Dark Reading’s report summarizes the trend.
That distinction matters. “More attackers” describes a change in the activity FireEye observed, not the full universe of exploitation. It also does not establish who discovered, developed or supplied any particular exploit. The practical lesson is narrower and more useful: defenders should not reserve attention to zero-day exploitation for a short list of elite actors.
What the evidence does—and doesn’t—show
The Dark Reading report describes an expansion in the attackers and zero-day vulnerabilities observed in FireEye’s analysis. It does not establish that use rose uniformly year by year, that every attacker type participated, or why the observed change occurred. A reported trend is not a complete map of the exploit market, and it cannot by itself identify the actor behind an incident.
A related FireEye Mandiant analysis provides context on patch timing. It examined a selected dataset of vulnerability and exploitation events spanning the first quarter of 2018 through the third quarter of 2019. Its statistics describe that dataset, not all vulnerabilities or attacks. The analysis reports that most vulnerabilities in its sample were zero-days and that 42 percent were exploited after a patch had been released. FireEye Mandiant’s analysis, hosted by Google Cloud, documents the study and its scope.
The same analysis reports an average of about nine days between disclosure and patch availability, with the average affected by an outlier: CVE-2019-0863 took five months from disclosure to patch. It also reports that a patch was available on the day of disclosure for 59 percent of vulnerabilities in the dataset. Among vulnerabilities first exploited after a patch, the study found that two were exploited within hours and 12 percent within a week. These are findings about the study’s selected observations, not universal rates or proof that releasing a patch causes an attack.
Why broader zero-day use changes the response
The operational lesson is straightforward: don’t make patch urgency depend on whether an incident appears to involve a state-backed group. FireEye’s reported 2019 finding suggests that zero-day exploitation should be considered across a wider range of attacker activity. The Mandiant study’s timing observations also underline the importance of acting when fixes become available. Together, they support a practical posture: track exposure, prioritize reachable and consequential systems, and reduce the time between a fix becoming available and its deployment.
This is not a call to treat every newly disclosed flaw as an emergency of equal weight. Teams have limited maintenance windows, complex dependencies and systems that cannot all be updated at once. Prioritize using evidence of exploitation, asset exposure and business impact rather than assumptions about an operator’s sophistication. The Mandiant analysis discusses factors attackers may weigh, including product use and exploit availability; any prevalence findings in that report apply to its dataset and period, not to the broader software landscape.
Turn the findings into patch decisions
Start with an inventory that shows where affected software runs. When a vendor releases a fix, identify exposed and business-critical instances first, then set a deployment window. If immediate patching is not feasible, document the temporary mitigation, its owner and a reassessment date. This is ordinary vulnerability management, but the study’s observations make “we’ll get to it next cycle” a decision worth scrutinizing.
Use threat intelligence to refine the queue, not replace it. Intelligence can help identify actively exploited vulnerabilities, relevant products and credible urgency. It cannot guarantee that an unobserved exploit is not in use, and an actor label rarely tells an administrator whether a vulnerable server is exposed. Prioritize what can be reached and what would matter if compromised; do not infer attribution from a vulnerability alone.
Finally, measure the gap between patch availability and deployment. An average can hide systems that remain exposed, and the Mandiant dataset includes both same-day patch availability and a long delay. Track exceptions, not just an organization-wide mean. If a system cannot be patched promptly, make the risk visible and revisit it as circumstances change.
The practical lesson: capability is not attribution
Zero-day use beyond elite APT groups widens the set of plausible threats; it does not tell us who used a particular flaw. Avoid both extremes: assuming only elite groups can exploit unknown vulnerabilities, and assuming the reported trend proves every attacker now has zero-day capability. FireEye’s observed expansion is a reason to consider the risk across threat models. The Mandiant study reinforces a practical response: patch promptly when risk warrants it, prioritize deliberately, and keep evidence about exploitation separate from attribution.
Sources: Dark Reading; FireEye Mandiant / Google Cloud.
zero-day exploits are no longer an elite-only capability