Skip to content
← Insights

Critical Vulnerabilities Get Patched in Hours. Period.

November 17, 2026 · 4 min read

AI Summary
Generating summary…

Your monthly patch cycle is a gift to attackers.

If a critical, internet-facing vulnerability drops this morning and your plan is to "get it into the next cycle," you've already decided to let someone else into your network for a few weeks. You just haven't admitted it yet.

The math doesn't work anymore

Synack's 2026 data puts the mean time from CVE publication to first observed exploitation at under 10 hours. In 2024 it was 56 days. Mandiant's M-Trends 2026 found 28.3% of CVEs exploited within 24 hours of disclosure. Rapid7 found 71% of tracked exploitation of high-severity bugs happens inside the first week.

Now look at our side. The 2026 Verizon DBIR puts median time to fully remediate known exploited vulnerabilities at 43 days, up from 32. By day seven, 60 to 70 percent of them are still open.

Attackers: hours. Us: six weeks, and getting slower.

That's not a gap. That's an open door with a welcome mat.

"But change control"

I've heard every objection. I've also cleaned up after every one of them.

"We need to test first." Test what, exactly? That the patch doesn't break the VPN that's about to be used to ransomware you? Smoke test one box, then go. You can roll back a bad patch. You can't roll back a domain admin compromise.

"We have a change window." Then you need an emergency change process that's pre-approved, with named people who can say yes at 2 a.m. If your CAB has to meet Thursday to approve a fix for something being exploited Monday, your CAB is part of the attack surface.

"It might cause downtime." It will. An hour of planned downtime beats three weeks of unplanned downtime while my team images your servers. I've watched companies sit offline for a month because nobody wanted to approve a 20-minute reboot.

"We're waiting for the vendor to confirm." Vendors are slow and sometimes wrong. If there's credible exploit code for a critical bug on something you expose, act on that.

Fear of downtime is understandable. It's also how a lot of the breaches I've worked got started.

What hours actually looks like

This isn't about heroics. It's about deciding in advance.

  • Know what you have before the advisory drops. If it takes two days to figure out whether you even run the affected product, you've already lost. Keep a live asset inventory with versions.
  • Define critical so nobody debates it. Internet-facing, remotely exploitable, on the CISA KEV list or with public exploit code. That's a drop-everything event. Not a ticket.
  • Pre-approve the emergency path. Write it down now. Who gets paged, who approves, what gets tested, how you roll back.
  • Set a clock. Target mitigation within hours of a critical, internet-facing disclosure. Track it like you track uptime.
  • If you can't patch, mitigate. Disable the vulnerable feature, pull the interface off the internet, apply the vendor workaround, push a WAF rule. Do something today.
  • Assume you were already hit. If the bug was exploited before you patched, patching doesn't evict anyone. Look for new accounts, web shells and odd logins going back to the disclosure date.
  • Report it up. Boards should see time-to-mitigate on critical exposures as a standing metric, right next to revenue.

Not everything is a fire

To be fair, not every CVE deserves this. More than 48,000 were published in 2025. Try to patch all of them in hours and you'll burn out your team and patch nothing well. The internal, authenticated, low-impact stuff can ride the normal cycle.

The critical, internet-facing ones can't. Those aren't maintenance. They're incident response that hasn't started yet.

Bottom line

Attackers are moving at machine speed. If your patching still moves at committee speed, you'll lose that race, on a timeline you could have predicted.

Hours. Not days. Not weeks. Not next cycle.

Next Step

Ready to strengthen your organization's resilience?

A 30-minute discovery call to discuss your cybersecurity posture, incident readiness, and whether advisory support is the right fit.