Cyber Attacks & Data Breaches: February 2026
A month-end digest of ransomware, data breaches, zero-days, supply chain compromises, state-linked activity and a dedicated look at AI-driven attacks.
Read this first: this page is compiled from public reports and trackers. Counts differ between sources (leak-site claims vs confirmed attacks), attributions are often preliminary, and claimed breaches may be unverified. Always verify with the original advisory before acting.
🔒Ransomware & Extortion
Qilin held the top spot for a second month, The Gentlemen nearly doubled and healthcare victims more than doubled.
VMware ESXi exploited by ransomware
CISA said ransomware operators were exploiting a critical ESXi sandbox-escape flaw to gain deep access to virtualized environments.
HypervisorSapienza University of Rome
One of Europe’s largest universities was knocked offline for days. The attack was linked to BabLock/Rorschach malware.
EducationBridgePay Network Solutions
A ransomware attack caused a payment platform outage across the US. Initial findings showed no payment card data compromised.
PaymentsConpet (Romania)
Qilin claimed nearly 1TB and posted samples; Conpet had not confirmed data loss.
ClaimedAdvantest and Washington Hotel (Japan)
Ransomware hit the chip test equipment maker, and the hotel group disclosed an attack on internal servers.
JapanUniversity of Mississippi Medical Center
A ransomware attack closed all 35 statewide clinic locations and cancelled appointments and elective surgeries.
HealthcareWynn Resorts
Confirmed a cybersecurity incident after a ransomware group listed the company on its leak site.
HospitalityLaw firms and manufacturers
INC listed Hawk Law Group and another firm, Play listed Deatak, and Everest claimed Iron Mountain.
Claimed🗄️Data Breaches & Leaks
ShinyHunters extortion and late-disclosed breaches produced the biggest numbers.
Amplifon
A breach at the hearing aid giant affected thousands of customers.
HealthcareOdido (Netherlands)
Attackers accessed a customer contact system and stole names, birth dates, bank account numbers and ID details of about 6.2 million customers.
TelecomFigure
ShinyHunters leaked about 2.5GB covering roughly 900,000 people.
FintechSubstack
Subscriber emails and phone numbers were exposed.
MediaCarGurus
ShinyHunters exposed account records of about 12.4 million users, forcing password resets.
ShinyHuntersConduent
State filings in February expanded the impact of the January 2025 attack to at least 25 million people.
Late disclosureCanada Goose (claimed)
ShinyHunters published data it attributed to Canada Goose, with over 600,000 customer records claimed.
ClaimedPanera Bread
A data breach hit the restaurant chain’s customers in early February.
Retail⚡Zero-Days & Exploits
Six exploited Microsoft zero-days, exploited management tools and a flood of KEV additions.
BeyondTrust RS and PRA
BeyondTrust auto-deployed a fix on Feb 2. CVE-2026-1731 was exploited within hours of a public proof of concept and added to KEV on Feb 13.
Remote accessMicrosoft Patch Tuesday
About 54 to 58 CVEs, including six zero-days exploited in the wild and three publicly disclosed. CISA added six to KEV the same day.
Patch TuesdayApple CVE-2026-20700
The first exploited Apple zero-day of 2026 allowed code execution via a memory flaw.
MobileCISA KEV additions
SolarWinds Web Help Desk (three-day deadline), Notepad++, Microsoft Configuration Manager and Apple flaws were added.
KEVChrome zero-day
Google fixed a flaw already exploited in the wild.
BrowserSmarterMail flaws
CVE-2026-24423 and CVE-2026-23760 were weaponized within days, with exploits and stolen admin credentials shared on Telegram.
Email serverCisco Catalyst SD-WAN
CVE-2026-20127 was added to KEV on Feb 25.
NetworkIvanti EPMM zero-days
CVE-2026-1281 and CVE-2026-1340 (CVSS 9.8) were exploited before the Jan 29 patch and drove several government breaches.
Zero-day🔗Supply Chain Attacks
Notepad++ update hijack
The maintainer disclosed a state-sponsored compromise of the update channel that lasted months. A suspected China-linked group, Lotus Blossom, delivered Cobalt Strike and the Chrysalis backdoor (CVE-2025-15556).
Update channelWealthsimple
The Canadian investment platform disclosed a breach through a third-party software package.
Software package🏛️State-Linked & Public Sector
Singapore telecoms
Singapore’s Cyber Security Agency revealed that China-linked UNC3886 had breached all four major telecom providers. Operation CYBER GUARDIAN took 11 months.
EspionageFrance bank account registry
A hacker used stolen credentials to access the national registry, exposing data tied to about 1.2 million accounts.
GovernmentEuropean Commission and Dutch bodies
Ivanti EPMM zero-days were used against the Commission (contained within nine hours) and the Dutch Data Protection Authority and Judicial Council.
GovernmentBadeSaba Calendar app
The popular Iranian prayer app was hacked to push anti-regime notifications during airstrikes on Iran.
Hacktivism🤖AI-Driven Cyber Attacks
February’s clearest case showed AI lowering the skill bar rather than inventing new exploits.
AI-augmented attacker breaches 600+ FortiGate firewalls
Amazon Threat Intelligence said a Russian-speaking, low-to-medium-skill actor used multiple commercial AI services to compromise over 600 devices in 55 countries between Jan 11 and Feb 18. No vulnerabilities were exploited; the attacker used exposed management ports and weak single-factor credentials.
AI-assistedAI-generated attack plans
The actor’s server held AI-written attack plans, AI-assisted parsing scripts and stolen firewall configs. Amazon said the code was fragile, typical of unrefined AI output.
AI toolingGemini abused across attack stages
Google reported threat actors abusing its Gemini model at every stage of attacks, mirroring Amazon’s findings.
Misuse reportOpenClaw agent warning
Microsoft warned against running OpenClaw AI agents on personal or enterprise workstations, advising dedicated virtual machines.
Agent risk- Lower barrier: a low-skill actor reached hundreds of firewalls with AI help and no exploits.
- Basics still win: exposed admin ports, weak passwords and missing MFA were the real entry points.
- Defense: never expose management interfaces, enforce MFA and isolate AI agents in virtual machines.
- CM-Alliance, Security Boulevard, SharkStriker, Bright Defense, Konbriefing and CSIS incident roundups
- Breachsense (ransomware counts); SANS, Recorded Future, Blackpoint, SC Media and Security Affairs (vulnerabilities and supply chain)
- Amazon Threat Intelligence, BleepingComputer, The Hacker News and Risky Bulletin (AI-driven attacks)
Independent roundup. Not affiliated with any vendor named. Next edition: March 2026.
