JPMorgan Chase, Visa and Mastercard have each confirmed security incidents at outside service providers after a data-extortion group posted what it says is their customers’ information online. Chase said on November 19, 2026, that attackers appear to have accessed marketing data held by vendors, which may include some customer names and email addresses. Visa and Mastercard said their own systems and payment networks were not affected. None of the three has named the vendor involved. Security researchers and trade press have, however, connected the leak to a campaign that abused trusted cloud access granted to a data analytics provider, and that link is where the lesson for every company with a vendor list sits.
TL;DR
- Chase, Visa and Mastercard confirmed incidents at third-party providers after a group calling itself n0n posted allegedly stolen customer data on November 16, 2026.
- The companies say no card numbers or payment systems were affected; the exposed data reportedly consists mainly of names, email addresses, some phone numbers and marketing records.
- n0n has claimed an attack on cloud analytics provider Qubole, whose AWS access to customer environments was abused between October 14 and 26, 2026, but the three companies have not confirmed which vendor was involved.

What each company has said
Chase posted its statement on X on the morning of November 19. “We have learned that third parties who provide services to the firm were the target of a cyber attack,” the bank said, adding that “based on our investigation so far, we believe the attackers accessed certain marketing data held by those vendors, which may include some customer names and email addresses.” According to SecurityWeek, the bank said there was no disruption to its services, that the incident was unrelated to problems some customers saw earlier in the week, and that it would contact affected customers and offer free credit monitoring where appropriate.
Visa told BleepingComputer that the claims are “associated with an incident involving a third-party service provider in Asia Pacific,” that Visa’s systems “were not involved and were not impacted,” and that no payments data such as card numbers was affected. Mastercard said it had been informed of an incident at a third-party provider that exposed “a limited amount of email addresses – non-payments information – from a small subset of cardholders enrolled in a specific benefit program.” It said its payment network was unaffected and the exposure had been contained.
That wording matters. Early summaries described Mastercard’s exposure as a “limited amount of information” from a subset of cardholders. The company’s statement to media is narrower: it describes email addresses only.
What the leaked files reportedly contain
The group calling itself n0n listed all three companies on its leak site on November 16 and marked them as not having paid, according to SecurityWeek, which reviewed the site. Reuters, as cited by SecurityWeek, found more than 16,000 customer email addresses in the JPMorgan-related files, some with phone numbers and postal codes, plus what appeared to be 124 email addresses belonging to bank employees. The Visa files reportedly held thousands of purported customer emails, phone numbers and marketing campaign records, mostly from Asia-Pacific markets. These are the group’s claims and journalists’ reviews of the files; the companies have not published customer counts.
The incident has been widely described as a ransomware attack. The public reporting so far describes data theft followed by extortion and publication. There are no reports of file encryption or system outages at the three companies, and each says its services were not disrupted. For notification analysis, that distinction is less important than what data was taken and whose systems held it.
The suspected vendor and the cloud trust problem
On November 11, n0n claimed responsibility for an attack on Qubole, a cloud data analytics platform owned by Idera. BleepingComputer reported that Qubole is suspected to be the provider the payment companies used; Qubole had not responded to its questions. The attack method was documented before the bank data appeared. In a November 9 analysis produced with Sentra and Mandiant, Wiz said a threat actor used IAM credentials from an AWS account associated with Qubole between October 14 and 26, 2026, to assume roles that customers had created to trust that account.
Wiz’s findings are uncomfortable reading for any third-party risk program:
- Most of the roles the attacker assumed were otherwise unused. In most environments Wiz observed, the attacker’s requests were the only role-assumption activity from the vendor account in October. The trust policies had simply never been removed.
- Every role-assumption request included the correct External ID and every one succeeded. Wiz recommends treating the External ID as non-secret, because it is logged in plain text and often stored in infrastructure code.
- Once inside, the attacker harvested secrets from AWS Secrets Manager, SSM Parameter Store and Lambda environment variables, then tested discovered AWS, Entra ID and Okta credentials.
- Data was taken opportunistically from whatever S3 buckets each role could reach, sometimes entire buckets at once.
- In several environments, missing S3 data-event logging meant responders could not tell which objects were read.
AWS told BleepingComputer its services “were not compromised” and that the activity involved credentials associated with a third-party provider. Wiz said a source-IP condition on the trust policies alone would have blocked every role assumption it observed. Wiz noted the n0n claim in an update but said it had not tied the campaign to other activity and had no direct insight into the actor’s motivation.
Vendor-originated breaches in financial services are not new. The SitusAMC breach showed how one service provider’s compromise can ripple across many lenders, and campaigns such as ShinyHunters’ extortion of ERP customers follow the same pattern of attacking one platform to reach many organizations.
Who has to notify whom
The answer depends on who the institution is and what data was exposed.
- Banks. National banks such as JPMorgan Chase Bank fall under the federal banking agencies’ computer-security incident notification rule, which requires notice to the primary regulator within 36 hours of determining that a “notification incident” has occurred. That term covers incidents that materially disrupt or degrade banking operations, so a vendor-held marketing data theft with no service disruption may not meet it. The rule separately requires bank service providers to notify their bank customers of certain incidents.
- Non-bank financial institutions. Companies subject to the FTC’s Safeguards Rule must notify the FTC no later than 30 days after discovering the unauthorized acquisition of unencrypted customer information involving at least 500 consumers. The rule also requires them to oversee service providers by selecting capable vendors, requiring safeguards by contract and periodically assessing them.
- State breach laws. Most state statutes are triggered by more sensitive data than names and email addresses, such as Social Security numbers, account numbers with access codes, or login credentials. Many also put the duty on the data owner even when a vendor held the data, with the vendor required to tell the owner.
- Outside the US. Visa’s provider is in Asia Pacific, so local breach notification regimes in that region may apply in addition to US rules.
Whether marketing data counts as “customer information” under the Gramm-Leach-Bliley Act is a real question rather than a technicality, since the fact that a person is a customer of a financial institution can itself be nonpublic personal information. Congress has periodically weighed changes to the GLBA framework, but the current rules are what apply here.
What vendor risk teams should do this week
- Inventory every cross-account cloud role. List each IAM role whose trust policy allows an outside account, and who approved it.
- Delete trust for vendors you no longer use. The roles abused here were mostly dormant.
- Pin trust to specific principals and networks. Replace account-wide trust with the vendor’s specific principal and add source-IP or VPC endpoint conditions.
- Cut what integration roles can reach. Vendor roles rarely need Secrets Manager, Lambda configuration or broad S3 read access.
- Turn on S3 data-event logging for sensitive buckets. Without it you cannot scope a breach, and you cannot narrow a notification.
- Put notification terms in the contract. Require vendors to report incidents within a defined window and to cooperate with forensics. Contractual data safeguards are only as useful as the reporting clause behind them.
- Tier vendors by access, not spend. A low-cost analytics tool with standing access to your cloud is a high-risk vendor. Sector-specific approaches to third-party risk management help set those tiers.
Frequently asked questions
Were card numbers stolen?
Visa and Mastercard say no payment data was affected, and Chase describes marketing data such as names and email addresses. Reporting on the leaked files describes emails, some phone numbers, postal codes and campaign records.
Has Qubole been confirmed as the vendor?
No. n0n claimed an attack on Qubole, and BleepingComputer reported Qubole is suspected to be the provider. None of the three companies has named its vendors, and Qubole had not commented publicly at the time of reporting.
What should affected customers watch for?
Leaked names, emails and phone numbers are useful for targeted phishing. Expect messages that reference a card benefit program or a bank relationship, and verify any request through the company’s official app or the number on the card.
Does a vendor breach shift legal responsibility to the vendor?
Generally not for notification. Under most US frameworks, the institution that owns the customer relationship remains responsible for notifying regulators and customers, even when the vendor held the data.
The dormant door
The detail that should stay with risk teams is not the bank names. It is that most of the cloud roles used in this campaign were not in active use. They had been created for an integration, left in place and forgotten. A vendor relationship does not end when the work stops. It ends when the access is removed.