Your Eyes in the DarkYour Eyes in the Dark — Proactive Intelligence

Ransomware/Knowledge

AI Fuels Surge in Ransomware Attacks, but Fewer Victims Are Paying

3 min read

Ransomware attacks are increasing as cybercriminals make more use of AI tools, but fewer organizations are paying ransoms, according to several reports released this month. Large language models (LLMs) and other AI systems let a growing number of ransomware groups write convincing phishing emails and find sensitive data quickly.

Attacks are up sharply worldwide
Organizations around the world are seeing far more ransomware incidents, while the number of attacks on US companies has stayed fairly steady. According to Halcyon:

  • Ransomware groups claimed 2,766 breaches in Q3 2026.
  • That’s up 31% from Q1 and 39% from Q2.

AI is removing the language barrier
Cynthia Kaiser, senior vice president of Halcyon’s Ransomware Research Center and a former deputy assistant director at the FBI’s Cyber Division, said AI is likely a major reason for the rise. LLMs’ ability to write convincing social-engineering messages, find sensitive data quickly, and assist attackers has largely driven the surge in attacks in other countries.

Kaiser said AI has removed the language barrier for phishing. Halcyon is now seeing much more convincing phishing in Japanese, for example, and countries like Japan are treating this almost as a state of emergency.

But fewer organizations are paying
More attacks have not necessarily meant more ransoms paid. Sophos’s The State of Ransomware in Retail 2026 report found:

Retail sector20252026
Attacks that encrypted data48%62%
Retailers that paid the ransom58%32%
  • 71% of retail firms recovered their data from backups.
  • Most recovered within a week.

Chester Wisniewski, global field CISO at Sophos, said other sectors, such as local governments and school districts, are also paying less often. According to him, they’re largely not paying, whether or not extortion or data theft is involved.

Ransom demands are falling
According to Sophos data, the median ransom demand has dropped:

YearMedian ransom demand
2024$2 million
2026$700,000

More ransomware groups than ever
The number of ransomware groups claiming victims grew from 75 in Q1 to 111 in Q3 2026, according to Halcyon. Wisniewski said he had seen five or six new names in just the last few weeks, a sign ransomware isn’t going away, even though it’s become less effective at making money for criminals.

Duy-Phuc Pham, a malware researcher at Trellix, said that at the end of the quarter, about a third of all claimed ransomware attacks came from groups that didn’t exist six months earlier.

According to Pham, The Gentlemen grew from a few dozen victims a year ago to become the most active group, ahead of Qilin. He said that success makes them an obvious next target for law enforcement, which spent this year going after the shared services these groups rely on, from criminal VPNs to laundering services.

AI enables faster attacks at greater scale
Pham said AI lets ransomware gangs expand their campaigns to more industries in more regions. That is part of why Europe has seen more attacks, while US victims’ share of incidents has stayed the same or fallen. AI helps attackers write convincing social-engineering messages in many languages, removing a major historical barrier to targeting smaller populations.

Pham said that when a tool can scan, exploit, and move through networks with little human effort, a small crew can do the work of a large one and hit far more victims, faster. That matches what Trellix sees on leak sites: more groups and more newcomers.

Manufacturing remains the top target
Kaiser said manufacturing is still the top target for ransomware gangs. Suppliers and partners often share the same platforms, so one compromise can expose engineering data beyond the targeted company. She expects manufacturing to remain the most targeted sector in 2027, especially as international organizations make up a growing share of victims.


Kaiser expects continued growth, especially outside the US. The big open question is whether more attacks will mean more money for attackers. She noted that countries newly feeling the impact are paying much more attention, which should improve international cooperation.

Get the latest dark web news in your inbox

Ransomware, breaches and exploited vulnerabilities, from The Darkweb News.

Free. No spam. Unsubscribe in one click.

Latest news

Ruby Patches “Ruby2Shell” Vulnerability That Could Lead to Remote Code Execution

Zero-Day / Patch

Fresh security updates for the Ruby programming language fix seven vulnerabilities in its regular expression engine and in string case conversion. Proof-of-concept (PoC) code demonstrating remote code execution (RCE) exists for five of them. The main flaw: Ruby2Shell (CVE-2026-132825) The top vulnerability, CVE-2026-132825 (CVSS 7.4), is a heap-based buffer overflow in string case conversion that could lead to RCE in certain scenarios. It’s been dubbed Ruby2Shell. How it works The flaw affects

China-Linked Armory Castle Exploits VMware vCenter Zero-Day for Admin Access and Data Theft

Zero-Day / Threat Intelligence

The suspected China-nexus cyber espionage group Armory Castle (formerly UNC5221) has been attributed to the zero-day exploitation of a critical flaw in VMware vCenter Server, starting in early September 2026. The vulnerability Tracked as CVE-2026-123890 (CVSS 9.8), the flaw is an authentication bypass in the VMware Directory Service. According to Google Threat Intelligence Group and Mandiant, it’s an unauthenticated Secure Remote Password (SRP-6a) session key disclosure vulnerability over DCE/R

New Unpatched Windows Zero-Day Lets Someone Run Programs on a Locked PC

Zero-Day

A newly disclosed, unpatched zero-day technique called FastUnlock lets someone with physical access to a locked Windows computer start programs without signing in. Whoever uses it gets only a limited account, not administrator control of the machine. Who disclosed it The method was posted by a researcher using the handles MSNightmare and Nightmare Eclipse. Endpoint-security firm ThreatLocker tested it the same day and confirmed it works. There is no fix from Microsoft yet, and nothing shows the