Snap published its second-half 2025 transparency report on 30 June 2026, covering 1 July to 31 December 2025. Snap said its Safety teams took 10,468,568 enforcements against 6,324,545 unique accounts. It reported a Violative View Rate of about 0.01% — one violating Snap or Story view per 10,000 — and 0.0003% for what it calls severe harms. In-app reports totalled 22,125,467, producing 6,858,386 reactive enforcements on 4,548,479 unique accounts, with a median turnaround of about one minute. Proactive automated tools accounted for 3,610,182 enforcements on 2,023,462 unique accounts.
Category counts are Snap’s taxonomy. Sexual content was 5,795,869 enforcements; child sexual exploitation and abuse 1,316,936; drugs 1,146,275; harassment 636,975; spam 347,196; hate speech 342,439; weapons 324,887; other regulated goods 275,820; threats and violence 197,967; self-harm and suicide 54,023; terrorism and violent extremism 22,264; false information 5,233; impersonation 2,684. Median times from detection to action were mostly one to seven minutes, with impersonation under one minute. Snap said excluded support-site and email-driven actions were less than 1% of volume.
CSEA-specific lines for H2 2025: 1,644,877 pieces of content enforced, 245,643 accounts disabled, 397,402 submissions to the US National Center for Missing and Exploited Children. Snap notes that one NCMEC submission can contain many files and that the number of media pieces submitted equals content enforced. Suicide-related resources were shared 38,503 times. Appeals of account locks: 471,892 total, 31,951 reinstated, 439,935 decisions upheld, median under one day. Ads reported 81,976 (removed-ad count appears in the H2 ads section snapshot as 81,976 reported; the loaded H2 extract listed reported ads without repeating the H1 removed-ad pair in the same way — this briefing therefore quotes H2 reported ads as 81,976 and does not invent an H2 removed-ad total).
Comparison with H1 2025, released 1 December 2025 and updated 19 May 2026, shows Snap itself restating CSEA numbers after a data-pull error: the H1 page says the correction added 32,809 NCMEC submissions and 104,833 content items. H1 totals were 9,674,414 enforcements and 5,794,201 unique accounts, with 19,766,324 in-app reports. Snap is therefore a source that revises. Readers should prefer the updated English page.
Geographic roll-ups for H2 put North America at 3,356,522 enforcements, Europe 3,082,639, rest of world 4,029,407. Unique-account sums do not add cleanly because Snap says a user’s region can change during the period. Bangladesh is not broken out in the narrative tables loaded here; country CSVs exist but were not fully ingested for this item.
Snapchat’s consumer product still differs from WhatsApp or iMessage on default message encryption: this transparency report is about Community Guidelines and legal requests, not a cryptography whitepaper. Enforcement statistics should not be read as a claim that disappearing Snaps are unrecoverable by Snap or by a forensic lab.
What this means in practice
Parents and schools can see that Snap says most sexual-content and CSEA actions now happen in minutes, and that it files hundreds of thousands of NCMEC reports in six months. That is Snap’s operational picture, not a guarantee a given report will be actioned in one minute. Users who appeal have, in Snap’s table, a small reinstatement fraction relative to upheld decisions. Cyber71 cannot compel a Snap reinstatement or obtain NCMEC packet contents.
What is still unclear
Violative View Rate methodology, hash-match false-positive rates, and how “sexual content” versus CSEA is labelled at the margin are Snap-defined. The H1 restatement shows that even published NCMEC counts can move. Country-level Bangladesh figures were not in the global HTML tables fetched. Encryption of Memories or My Eyes Only is outside this report’s scope.
Snap’s own year-on-year story in these two reports is faster median review and more proactive CSEA and drugs detections, alongside policy reclassifications that moved some sexual-harassment volume between buckets in H1. That is why this briefing quotes both editions rather than a single headline. Anyone citing Snap in a Bangladesh safety debate should attribute the 0.01% VVR and the NCMEC counts to Snap’s measurement, note the May 2026 H1 restatement, and avoid converting enforcements into a prevalence of harm among Bangladeshi teens — Snap did not publish that rate in the global HTML.
Sources
- Snap Inc. — Snapchat Transparency Report, July–December 20252026-09-24
- Snap Inc. — Snapchat Transparency Report, January–June 20252026-09-24
Cyber71 links the primary material behind each report so readers can check it directly. External pages are not under our control and may change.
