As reported by BleepingComputer, J.P. Morgan Chase, Visa, and Mastercard each disclosed customer data exposure stemming from a cyberattack at a shared third-party service provider—widely believed to be Qubole, a cloud data analytics platform owned by Idera. The n0n ransomware group has claimed responsibility, asserting they used compromised Qubole credentials to pivot into customer AWS environments via pre-configured IAM roles, exfiltrating nearly 54 GB of data across more than 30,000 files.

Cloud Security Alert: Morgan Chase, Visa, and Mastercard each disclosed customer data exposure stemming from a cyberattack at a shared third-party service provider—widely believed to be Qubole, a cloud data analytics platform owned by Idera.

The Real Story: IAM Roles as a Lateral Movement Vector

While the headlines focus on which brand-name financial institutions were affected, the technical reality is more instructive. The attackers didn't breach Chase, Visa, or Mastercard directly. They breached a vendor that those organizations had trusted with cloud-native access—specifically, IAM roles that allowed Qubole to read from and interact with customer AWS environments.

This is the supply chain risk pattern that cloud security teams have been warning about for years, and it remains alarmingly common. Organizations grant analytics platforms, data processing vendors, and SaaS integrations persistent IAM roles that are often over-scoped at creation time and never revisited. When those vendor credentials are compromised, the attacker inherits whatever access the role provides—no additional exploitation required.

The attack surface isn't your vendor. It's the IAM role you gave your vendor.

Who Is at Risk

Any organization that has granted a third-party platform persistent access to its cloud environment through IAM roles, service accounts, or long-lived API keys is exposed to this same pattern. The financial services sector is particularly vulnerable because of the density of analytics, reporting, and compliance vendors embedded in their data pipelines. But the same architecture exists across healthcare, retail, manufacturing, and technology sectors.

The 30,743 files allegedly exfiltrated from Qubole reportedly included customer contracts, pricing records, and staff information—data that is valuable for follow-on social engineering, competitive intelligence, or extortion leverage. The fact that n0n ransomware claimed the attack suggests dual monetization: extortion against Qubole directly and potential secondary use of stolen customer data.

Why This Keeps Happening

Who Is at Risk
Persistent over-privileged roles: IAM roles granted for integration are rarely scoped to true least privilege. They often include broad S3 read access, list permissions across all buckets, or unnecessary cross-service permissions.
Credential lifecycle gaps: Vendor credentials are frequently long-lived, shared across environments, and not rotated on any meaningful cadence.
Visibility blind spots: CloudTrail may log the API calls, but if no one is correlating unusual data access patterns against a baseline, exfiltration of 54 GB can proceed undetected for weeks.
Trust-by-default: Onboarding a vendor often means granting access first and assessing scope later—if at all.

Shield53 Recommendations

Immediate Actions

  • Audit all third-party IAM roles: Inventory every IAM role, service account, and access key granted to external vendors. For each, document what data it can reach and when it was last reviewed.
  • Reduce role scope: Replace broad S3 access with bucket-specific and prefix-specific policies. Remove ListBucket permissions where only specific object GETs are needed.
  • Rotate vendor credentials: Force rotation of all long-lived access keys and IAM role trust policies associated with third-party integrations. Move to temporary credentials via STS AssumeRole where possible.
  • Enable data exfiltration detection: Deploy CloudTrail + GuardDuty or equivalent monitoring tuned for anomalous S3 GetObject volume, unusual source IP ranges, and off-hours access patterns.

Strategic Hardening

  • Implement just-in-time access: Replace persistent vendor roles with time-bound, approval-gated access using tools like AWS IAM Identity Center or third-party JIT platforms.
  • Adopt a vendor cloud access policy: Require security review of any IAM role or access policy before a vendor integration goes live. Include data classification—no vendor role should touch sensitive data stores without explicit authorization.
  • Network-level isolation: Where feasible, restrict vendor access to specific VPC endpoints or private subnets rather than allowing internet-facing API access.
  • Quarterly access reviews: Schedule mandatory reviews of all third-party cloud access. Remove roles for vendors no longer in active use.

The Qubole incident is not an anomaly—it is a high-profile instance of a systemic pattern. Every cloud-enabled organization should treat third-party IAM roles as privileged access, governed with the same rigor as internal administrator credentials. The cost of doing otherwise is visible in today's headlines.