Capcom
No passwords, no payment cards. Just a precise record of when 11.7 million people were at their computers.
What happened?
Capcom confirmed on 19 October 2026 that a breach of its Capcom ID account system exposed records for approximately 11.7 million accounts. Capcom ID is the single sign on that links a player across the publisher's titles and services.
The company acknowledged the incident publicly in the days that followed, and reporting through 22 and 23 October refined the record count and the field list.
What data was actually inside?
Email addresses, account creation dates, last login timestamps, login counts and IP addresses. A subset of records also included dates of birth, gender, nicknames and countries.
What is absent is as notable as what is present. No passwords in the reporting, no payment card data. By the standard checklist this is a low severity breach, and by the standard checklist the standard checklist is wrong here, because the fields that were taken are behavioural rather than identifying.
Who gets hurt and how?
Players, including a meaningful number who created an account years ago for one game and have not thought about it since.
An IP address paired with a last login timestamp and a login count is a location and a schedule. It places a person in a city, often in a neighbourhood, and it describes when they are at the device. Combine that with a date of birth and a nickname and the record stops being an account and becomes a profile. For anyone with a reason to avoid being located, which includes people under harassment campaigns, that combination is a safety matter and not a fraud matter.
The email address does the conventional work. Eleven point seven million addresses verified as belonging to active gaming accounts is a high quality phishing list, and the account creation date tells the sender which era of game to reference to be believed.
What did they think they were doing right?
Capcom separated the sensitive material from the account layer. Passwords and payment data do not appear in this incident, which indicates they were held differently and protected accordingly. That is the correct design and it worked.
The assumption left standing is that an account record minus the credentials and the card is a low value object. It is a reasonable assumption about any single field and a poor one about the set. Login telemetry exists for fraud detection and capacity planning, it accumulates quietly because it is cheap to store, and it is almost never classified as personal data in the way an address is.
What did they not know about their own data?
How much behavioural history had collected in the identity system. An account created in 2016 with a login count attached has a decade of implied activity, and nobody at any point decided that this was worth retaining for a decade. It was retained because deleting it required a decision and keeping it did not.
The IP address is the field most likely to have been invisible on an internal data map. It is captured for abuse prevention, it is treated as an operational log rather than a personal attribute, and under GDPR it is personal data. A publisher with European players therefore held regulated data in a field its own classification probably listed as telemetry.
The breach did not reveal what Capcom stored. It revealed what Capcom had not described to itself.
If your business runs on databases, you probably have similar records, customer data, credentials, financial information. Do you know what's actually in yours?
Not sure where to start? How to find sensitive data across your estate
What does attribution look like the morning after?
No threat actor has been named in the available reporting.
Notification for 11.7 million gaming accounts crosses every jurisdiction the publisher sells into, and the obligations diverge sharply. An IP address plus a date of birth triggers a GDPR assessment in the European Union that it does not trigger in several other markets. The operational burden is therefore not one notification exercise but a dozen, and the determining question in each is which fields a given regulator considers personal, which is a question about the field list rather than about the attack. The ASOS entry in this series shows the same multi jurisdiction problem in retail.
What would have changed the outcome?
Treating login telemetry as personal data at the point it was collected rather than at the point it was stolen. IP addresses and timestamps are regulated in the markets Capcom sells into, and the classification that governs them should have matched that before an incident forced the question.
Second, expiring it. A login count that spans ten years is ten years of exposure in exchange for an analytics value that decays within weeks. Almost none of what was taken here needed to still exist.
Capcom found out the hard way.
Your team could spend the next 6 months rebuilding systems, notifying customers, and answering legal questions. Or you could spend 24 hours finding out what's actually at risk.
Want to understand the process first? Read our guide to sensitive data discovery .