Ransomware Intelligence / Groups / akira

Ransomware group

akira

Akira is a ransomware group first observed in March 2023, targeting both Windows and Linux environments, with a particular focus on corporate networks and VMware ESXi servers. The group employs a double extortion model, stealing sensitive data before encrypting systems and threatening to leak it on a Tor-based leak site if ransom demands are not met. Akira typically gains initial access through exploitation of unpatched VPN services, compromised RDP credentials, phishing, or abuse of legitimate remote administration tools. Its Windows variant uses the Windows CryptoAPI to encrypt files, appending the “.akira” extension while skipping critical system folders to maintain system stability. Ransom demands have ranged from $200,000 to over $4 million, typically requested in Bitcoin, and the group has been linked to high-profile incidents affecting education, manufacturing, and healthcare sectors. Akira appears to operate independently rather than as a Ransomware-as-a-Service, and continues to evolve, with recent variants improving encryption speed and evasion techniques.

  • 12claims, 7 days
  • 38claims, 30 days
  • 95claims, 90 days
  • 2026-07-17first seen
  • 2026-10-15last seen

These are claims posted by the group on its leak site, counted when we first see them. They are not confirmed incidents, and we never name a victim.

Activity, last 26 weeks

New leak-site claims per week.

Week of 2026-04-20: 0Week of 2026-04-27: 0Week of 2026-05-04: 0Week of 2026-05-11: 0Week of 2026-05-18: 0Week of 2026-05-25: 0Week of 2026-06-01: 0Week of 2026-06-08: 0Week of 2026-06-15: 0Week of 2026-06-22: 0Week of 2026-06-29: 0Week of 2026-07-06: 0Week of 2026-07-13: 2Week of 2026-07-20: 7Week of 2026-07-27: 2Week of 2026-08-03: 4Week of 2026-08-10: 6Week of 2026-08-17: 5Week of 2026-08-24: 12Week of 2026-08-31: 11Week of 2026-09-07: 7Week of 2026-09-14: 10Week of 2026-09-21: 9Week of 2026-09-28: 8Week of 2026-10-05: 5Week of 2026-10-12: 9
Claims per week, oldest to newest. Each bar is a full Monday-to-Sunday week; the newest bar is the current week so far.

Countries claimed, last 90 days

  1. United States 46
  2. Unknown 27
  3. Germany 5
  4. United Kingdom 5
  5. Canada 3
  6. Brazil 2
  7. Spain 2
  8. United Arab Emirates 1
  9. Czechia 1
  10. Italy 1
  11. Norway 1
  12. SI 1

Vulnerabilities with known ransomware use

From CISA's Known Exploited Vulnerabilities catalog, newest first. The catalog does not attribute entries to a specific group; this is the current patch list for any ransomware operator.

Added CVE Vendor / product Vulnerability
2026-08-18 CVE-2026-59310 Broadcom VMware vCenter Broadcom VMware vCenter Path Traversal Vulnerability
2026-08-05 CVE-2026-63077 JetBrains TeamCity JetBrains TeamCity Deserialization of Untrusted Data Vulnerability
2026-07-29 CVE-2026-20316 Cisco Secure Firewall Management Center (FMC) Cisco Secure Firewall Management Center Use of Hard-coded Password Vulnerability
2026-07-14 CVE-2026-15410 SonicWall SMA1000 Appliances SonicWall SMA1000 Appliances Code Injection Vulnerability
2026-07-14 CVE-2026-15409 SonicWall SMA1000 Appliances SonicWall SMA1000 Appliances Server-Side Request Forgery Vulnerability
2026-07-01 CVE-2026-45659 Microsoft SharePoint Server Microsoft SharePoint Server Deserialization of Untrusted Data Vulnerability
2026-06-25 CVE-2026-12569 PTC Windchill and FlexPLM PTC Windchill and FlexPLM Improper Input Validation Vulnerability
2026-06-12 CVE-2026-35273 Oracle PeopleSoft Enterprise PeopleTools Oracle PeopleSoft Enterprise PeopleTools Missing Authentication for Critical Function Vulnerability

Data sources

  • Source: CISA KEV
    updated 2 hours ago
  • Opening demands are read from leaked ransomware negotiation chats collected by Ransomchats (MIT licence), using an OpenAI language model. These chats are not a representative sample.
    updated 2 hours ago
  • Source: RansomLook
    updated 2 hours ago
  • Sector, country and region labels are inferred by SecuritySnares from the published organisation name using an OpenAI language model; treat them as estimates.
    updated 2 hours ago

Leak-site listings reflect claims by threat actors and are not independently verified. Victims are never named on this site unless they have disclosed the incident themselves. Read the methodology.