Back to Intelligence

Unlimited Technology Systems Breach: 3.8 Million Healthcare Records Exfiltrated From Commercial Data Center — Detection and Hardening Guide

SA
Security Arsenal Team
August 9, 2026
11 min read

Unlimited Technology Systems, a U.S.-based healthcare technology company, has disclosed a data breach affecting more than 3.8 million individuals. According to the disclosure, unauthorized actors gained access to one of the company's commercial data centers during a six-day window between October 5 and October 10, 2025, and exfiltrated personal, medical, and insurance data belonging to healthcare patients.

This is a textbook example of the threat profile healthcare technology providers face in 2026: attackers are not just going after hospitals — they are going after the infrastructure providers that aggregate data across multiple covered entities. A single compromise at a technology vendor cascades into a multi-client, multi-state breach notification event with HIPAA, HHS OCR reporting, and state attorney general implications.

Six days of dwell time inside a data center environment means the attackers had ample opportunity to stage, compress, and exfiltrate data at scale. If your SOC is not instrumented to catch bulk data staging and anomalous outbound transfer from server segments, this same scenario plays out silently in your environment. This post breaks down the defensive lessons, provides detection content for the behaviors that almost certainly occurred during this intrusion, and gives you a hardening checklist grounded in what actually failed here.

Technical Analysis

What We Know

  • Victim: Unlimited Technology Systems, a U.S. healthcare technology company
  • Attack window: October 5–10, 2025 (six days of unauthorized access)
  • Target: A commercial data center operated by/used by the company
  • Data stolen: Personal identifiable information (PII), medical data (PHI), and insurance information for 3.8+ million individuals
  • Disclosure: Public breach notification, with regulatory notification obligations under HIPAA (HHS OCR reporting required for breaches affecting 500+ individuals)

No CVE has been publicly attributed to this intrusion, and no specific initial access vector has been disclosed. That is typical for breach disclosures of this kind — and it is also a signal. When there is no single vulnerability to point to, the root cause is almost always one or more of the following, in order of prevalence in our IR engagements involving healthcare data centers:

  1. Compromised credentials for remote access infrastructure (VPN, RDP, VDI, or a vendor maintenance tunnel) without MFA or with MFA fatigue
  2. Exposed management interfaces — out-of-band management (iLO/iDRAC/IPMI), hypervisor consoles (vCenter/ESXi), or backup infrastructure consoles reachable from less-trusted segments
  3. Flat network architecture inside the data center, allowing an initial foothold to reach database and file storage tiers without further authentication
  4. Inadequate egress monitoring, allowing multi-terabyte exfiltration to cloud storage or attacker-controlled infrastructure without triggering alerts

The Likely Attack Chain

Based on the six-day dwell time and the volume of data involved (3.8M records implies bulk database or file-share access), defenders should model this intrusion against the following chain:

  • Initial Access (TA0001): Valid accounts (T1078) against internet-facing remote access, or exploitation of an exposed appliance
  • Discovery (TA0007): Network share discovery (T1135), enumeration of database servers and file shares hosting PHI
  • Collection (TA0009): Data from network shares (T1039) and databases; archive collected data (T1560) — typically with 7-Zip, WinRAR, or native tar/zip on Linux storage nodes
  • Exfiltration (TA0010): Exfiltration over web services (T1567) to attacker-controlled cloud storage (MEGA, Backblaze, S3-compatible endpoints) or direct transfer over HTTPS/SFTP; increasingly via Rclone, which appears in the majority of large-scale data theft cases we respond to

Severity and Regulatory Exposure

With 3.8 million affected individuals, this breach sits well above the HHS OCR "wall of shame" threshold and will trigger:

  • Mandatory HIPAA breach notification to HHS, affected individuals, and likely media (45 CFR §§ 164.400–414)
  • State-level notifications across most U.S. jurisdictions (insurance data implicates state insurance regulators as well)
  • Near-certain OCR investigation into Security Rule compliance (risk analysis, access controls, audit controls, transmission security)
  • Class-action litigation exposure, which now routinely exceeds $1,000 per record in healthcare settlements

The defenders' takeaway: the difference between a containable incident and a 3.8M-record breach is egress detection and response speed. Six days is a long time to miss bulk outbound transfer.

Detection & Response

The detections below target the behaviors that are near-universal in data-center PHI theft cases: bulk archive creation on server systems, exfiltration tooling execution, and anomalous outbound transfer volumes. They are tuned for server segments and should not fire broadly in well-baselined environments.

YAML
---
title: Archive Utility Execution on Server Systems - Potential Data Staging
id: 3f8a2c91-7b4e-4d1a-9c55-8e2f6a0b1d34
status: experimental
description: Detects execution of archive/compression utilities with command lines consistent with bulk data staging on Windows server systems, a common precursor to exfiltration in data center intrusions.
references:
  - https://attack.mitre.org/techniques/T1560/001/
  - https://securityaffairs.com/196843/data-breach/unlimited-technology-systems-data-breach-exposes-data-of-3-8-million-healthcare-patients.html
author: Security Arsenal
date: 2026/01/15
tags:
  - attack.collection
  - attack.t1560.001
logsource:
  category: process_creation
  product: windows
detection:
  selection_img:
    Image|endswith:
      - '\7z.exe'
      - '\7za.exe'
      - '\rar.exe'
      - '\winrar.exe'
  selection_cli:
    CommandLine|contains:
      - ' a '
      - ' -r'
      - ' -p'
  filter_backup_paths:
    CommandLine|contains:
      - '\backup\'
      - '\veeam\'
      - '\sccm\'
  condition: selection_img and selection_cli and not filter_backup_paths
falsepositives:
  - Legitimate backup and software packaging operations; baseline per-server and alert on deviation
level: high
---
title: Rclone or Cloud Exfiltration Tool Execution
id: 9c1d5e47-2a8f-4b36-8d70-4f3a9b5c2e81
status: experimental
description: Detects execution of Rclone or similar cloud sync tooling frequently abused for bulk data exfiltration to attacker-controlled storage, as seen in healthcare data theft intrusions.
references:
  - https://attack.mitre.org/techniques/T1567/002/
  - https://securityaffairs.com/196843/data-breach/unlimited-technology-systems-data-breach-exposes-data-of-3-8-million-healthcare-patients.html
author: Security Arsenal
date: 2026/01/15
tags:
  - attack.exfiltration
  - attack.t1567.002
logsource:
  category: process_creation
  product: windows
detection:
  selection_img:
    Image|endswith:
      - '\rclone.exe'
      - '\megacmd.exe'
      - '\MEGAcmd.exe'
  selection_name:
    OriginalFileName:
      - 'rclone.exe'
  condition: selection_img or selection_name
falsepositives:
  - Approved cloud backup tooling using rclone; restrict to an allowlist of service accounts and paths
level: high
---
title: Mass Outbound Data Transfer From Server Segment
id: 5b7e3f09-1c6a-4d82-9e14-7c8a5d0f3b62
status: experimental
description: Detects unusually large outbound network connections from database or file server systems to non-corporate destinations, indicative of bulk data exfiltration from a data center environment.
references:
  - https://attack.mitre.org/techniques/T1041/
  - https://securityaffairs.com/196843/data-breach/unlimited-technology-systems-data-breach-exposes-data-of-3-8-million-healthcare-patients.html
author: Security Arsenal
date: 2026/01/15
tags:
  - attack.exfiltration
  - attack.t1041
logsource:
  category: network_connection
  product: windows
detection:
  selection:
    DestinationIp|cidr:
      - '0.0.0.0/0'
    Initiated: 'true'
  filter_rfc1918:
    DestinationIp|cidr:
      - '10.0.0.0/8'
      - '172.16.0.0/12'
      - '192.168.0.0/16'
  filter_known_services:
    DestinationHostname|endswith:
      - '.microsoft.com'
      - '.windowsupdate.com'
      - '.digicert.com'
  condition: selection and not filter_rfc1918 and not filter_known_services
falsepositives:
  - Cloud replication and SaaS integrations; pair with byte-volume aggregation in your SIEM and alert on per-host daily egress exceeding baseline
level: medium
KQL — Microsoft Sentinel / Defender
// Hunt: Servers with abnormal daily egress volume (potential bulk exfiltration)
// Baseline per-device outbound bytes over 7 days, flag hosts exceeding 3x their median
let Lookback = 7d;
let EgressData =
    DeviceNetworkEvents
    | where TimeGenerated > ago(Lookback)
    | where ActionType == "ConnectionSuccess"
    | extend RemoteIPType = iff(ipv4_is_private(RemoteIP), "private", "public")
    | where RemoteIPType == "public"
    | summarize DailyBytes = sum(BytesSent) by DeviceName, bin(TimeGenerated, 1d);
let Baseline =
    EgressData
    | summarize MedianDailyBytes = percentile(DailyBytes, 50) by DeviceName;
EgressData
| join kind=inner Baseline on DeviceName
| where DailyBytes > MedianDailyBytes * 3 and DailyBytes > 10GB
| project TimeGenerated, DeviceName, DailyBytes, MedianDailyBytes
| order by DailyBytes desc;

// Hunt: Execution of exfiltration tooling on server operating systems
DeviceProcessEvents
| where TimeGenerated > ago(14d)
| where FileName in~ ("rclone.exe", "megacmd.exe", "7z.exe", "7za.exe", "rar.exe", "winscp.exe", "filezilla.exe")
   or ProcessCommandLine has_any ("rclone", "mega.nz", "copy --", "sync --")
| where DeviceName has_any ("srv", "db", "sql", "fs", "dc") // scope to server naming convention
| project TimeGenerated, DeviceName, AccountName, FileName, ProcessCommandLine, InitiatingProcessFileName, InitiatingProcessAccountName
| order by TimeGenerated desc;

// Hunt: New outbound destinations never seen before from server segments (30-day first-seen)
let KnownDest =
    DeviceNetworkEvents
    | where TimeGenerated between (ago(30d) .. ago(1d))
    | where ActionType == "ConnectionSuccess"
    | summarize by RemoteIP;
DeviceNetworkEvents
| where TimeGenerated > ago(1d)
| where ActionType == "ConnectionSuccess"
| where not(ipv4_is_private(RemoteIP))
| where RemoteIP !in (KnownDest)
| summarize Connections = count(), BytesSent = sum(BytesSent), Processes = make_set(FileName) by DeviceName, RemoteIP, RemoteUrl
| where BytesSent > 100MB
| order by BytesSent desc;
VQL — Velociraptor
-- Hunt: exfiltration tooling and archive staging artifacts on endpoints
-- Scope: run against server collections during IR or proactive healthcare environment sweeps
SELECT Pid,
       Name,
       Exe,
       CommandLine,
       Username,
       CreateTime
FROM pslist()
WHERE Exe =~ '(?i)(rclone|megacmd|7z|7za|rar|winrar|winscp|filezilla)'
   OR CommandLine =~ '(?i)(rclone|mega\.nz|--password|-p[A-Za-z0-9])'

-- Correlate with large recently-modified archives in common staging directories
SELECT FullPath,
       Size,
       Mtime,
       Btime
FROM glob(globs=['C:/Users/*/AppData/Local/Temp/*.zip',
                 'C:/Users/*/AppData/Local/Temp/*.rar',
                 'C:/Users/*/AppData/Local/Temp/*.7z',
                 'C:/ProgramData/*.zip',
                 'C:/ProgramData/*.rar',
                 'C:/ProgramData/*.7z',
                 'C:/Windows/Temp/*.7z'])
WHERE Size > 50000000
  AND Mtime > now() - 1209600
ORDER BY Mtime DESC
PowerShell
# Unlimited Technology Systems-style breach: server egress audit and exfil-tool hardening
# Run on data center jump hosts / management servers, or deploy via GPO/Intune to server OUs

# 1. Inventory installed/running exfiltration tooling across the local system
$suspectTools = @('rclone','megacmd','winscp','filezilla','psftp','curl')
Get-Process | Where-Object { $_.Name -in $suspectTools } |
  Select-Object Name, Id, Path, StartTime | Format-Table -AutoSize

# 2. Review outbound connections from server processes over the last boot session
Get-NetTCPConnection -State Established |
  Where-Object { $_.RemoteAddress -notmatch '^(10\.|172\.(1[6-9]|2[0-9]|3[0-1])\.|192\.168\.|127\.|169\.254\.)' } |
  ForEach-Object {
    $proc = Get-Process -Id $_.OwningProcess -ErrorAction SilentlyContinue
    [PSCustomObject]@{
      RemoteIP  = $_.RemoteAddress
      RemotePort= $_.RemotePort
      Process   = $proc.Name
      ProcPath  = $proc.Path
    }
  } | Sort-Object Process -Unique | Format-Table -AutoSize

# 3. Hunt recent archive staging in common drop locations (last 14 days, >50MB)
$cutoff = (Get-Date).AddDays(-14)
$paths = @("$env:TEMP", 'C:\ProgramData', 'C:\Windows\Temp')
foreach ($p in $paths) {
  Get-ChildItem -Path $p -Recurse -Include *.zip,*.rar,*.7z -ErrorAction SilentlyContinue |
    Where-Object { $_.Length -gt 50MB -and $_.LastWriteTime -gt $cutoff } |
    Select-Object FullName, @{N='SizeMB';E={[math]::Round($_.Length/1MB,2)}}, LastWriteTime
}

# 4. Verify Windows Defender Firewall egress logging is enabled on server profile
Get-NetFirewallProfile -Profile Domain,Private |
  Select-Object Name, LogAllowed, LogBlocked, LogFileName | Format-Table -AutoSize

# 5. Enable outbound connection auditing (Success + Failure) for exfil detection pipelines
auditpol /set /subcategory:"Filtering Platform Connection" /success:enable /failure:enable
auditpol /get /subcategory:"Filtering Platform Connection"

Remediation and Hardening

If you are a customer or downstream covered entity of Unlimited Technology Systems, treat this as a supply-chain incident: activate your third-party incident response playbook, request the forensic findings (initial access vector, dwell-time timeline, exact data elements), and prepare your own notification obligations if your patients' data was in scope. If you operate a comparable healthcare data center, the following hardening measures directly address the failure modes this breach exposes:

  1. Egress control is the highest-leverage fix. Database and file server tiers should have no direct internet egress. Route all outbound traffic through an explicit proxy with TLS inspection where legally/operationally feasible, and deny-by-default destination categories like personal cloud storage, file-sharing, and newly-registered domains. A 3.8M-record exfiltration cannot complete if servers cannot talk to the internet.
  2. Baseline and alert on egress volume per host. The single most reliable detection in this breach class is volumetric: a file server that historically sends 2 GB/day suddenly sending 400 GB is not a tuning problem, it is an incident. Implement the KQL logic above as an analytic rule with per-device baselining.
  3. Enforce MFA and conditional access on every remote path into the data center — VPN, RDP gateways, hypervisor management (vCenter/ESXi), out-of-band management (iLO/iDRAC), and backup consoles. Vendor maintenance tunnels must be just-in-time, logged, and time-boxed.
  4. Segment ruthlessly. Management interfaces, backup infrastructure, and database tiers must be isolated from general server VLANs. Attackers who land in a flat data center network reach PHI storage in hours; microsegmentation turns that into a detection opportunity.
  5. Protect the backup and storage control planes. In healthcare breaches we consistently see attackers enumerate or tamper with backup consoles to maximize leverage. Backup admin interfaces should be on dedicated jump hosts with PAW-level controls.
  6. Database activity monitoring (DAM) on PHI stores. Alert on bulk SELECT/export operations, new service accounts touching patient tables, and access volume anomalies per principal. The Collection phase is your last in-network tripwire before data leaves.
  7. Validate HIPAA Security Rule audit controls (45 CFR §164.312(b)). Six days of undetected access strongly suggests audit logs either were not centralized or were not being reviewed. Ship server, authentication, and network flow logs to a SIEM with retention sufficient to reconstruct a full intrusion timeline — this is also what OCR will ask for.
  8. Tabletop the notification clock. HIPAA requires notification without unreasonable delay and no later than 60 days from discovery; state laws (e.g., 30 days in several states) are tighter. Pre-draft notification templates, credit-monitoring vendor contracts, and counsel engagement before you need them.

For affected individuals, the standard guidance applies: enroll in offered credit/identity monitoring, watch for Explanation of Benefits (EOB) anomalies (medical identity theft is the primary downstream risk of PHI theft, far more than financial fraud), and place fraud alerts with the bureaus if SSNs were involved.

Related Resources

Security Arsenal Healthcare Cybersecurity AlertMonitor Platform Book a SOC Assessment healthcare Intel Hub

Is your security operations ready?

Get a free SOC assessment or see how AlertMonitor cuts through alert noise with automated triage.

Unlimited Technology Systems Breach: 3.8 Million Healthcare Records Exfiltrated From Commercial Data Center — Detection and Hardening Guide | Security Arsenal | Security Arsenal