Qilin leaks hundreds of gigabytes from Italian telecom and cloud operator Retelit

Italian telecommunications and cloud operator Retelit was hit by the Qilin ransomware group, which published a large cache of stolen internal files before the company had said anything publicly. The incident came to light on 4 August 2026 through an investigation by Italian outlet IrpiMedia. Retelit’s multicloud service is qualified for use by Italian public administrations.

What happened

  • Qilin added Retelit S.p.A. to its leak site on 11 July 2026 and posted a sample of internal documents and passports on 14 July.
  • Between 30 July and 1 August the group published the bulk of the material. Its leak page lists about 270,000 files, which IrpiMedia estimates at no less than 300 GB. The release suggests that no ransom was paid.
  • The leaked material reportedly includes identity documents, supplier lists, management information, orders and customer project files.
  • On 30 July, Italy’s CERT-AGID began alerting security contacts at public administrations that might be affected.

In a right-of-reply statement, Retelit said the attack took place on 8 June. It affected 3 of its 38 national data centres, roughly 7% of hosted systems, and did not touch connectivity services. The company acknowledged evidence consistent with data exfiltration. It said it had notified customers, the national cybersecurity agency ACN, the CSIRT, the Polizia Postale and the data protection authority, and had set up a war room with forensic specialists. IrpiMedia identified the affected sites as Verona, Rome and Milan, the last of them an ACN-certified backup site.

Impact

Retelit’s customers reportedly include defence group Leonardo, at least three digital identity providers and 193 public administrations. CERT-AGID’s notice named Cineca, Lepida and InfoCamere, which provide document preservation, SPID and digital signature services. A customer quoted by IrpiMedia said only about 70% of a backup could be recovered, and another said it permanently lost data hosted in the affected facility. According to an unconfirmed account, the intrusion started from a system administrator’s workstation and moved to the virtualisation layer using stolen credentials.

Recommendations

  • Customers of affected providers should ask for a concrete list of impacted services and data, and should rotate any credentials or API keys stored on hosted systems.
  • Cloud operators should isolate administrator workstations, use privileged-access workstations and protect hypervisor management with phishing-resistant MFA.
  • Keep backup infrastructure logically separate from production credentials so one compromise cannot reach both.
  • Check whether administrator credentials appear in infostealer logs; services such as TAYNOPIS can help identify exposed accounts.

Sources