Bloomberg Investigation (Baker and Bengani) Finds Matryoshka Bots Inflating View Counts on Fake Videos Against Democrats in 20 Senate and 29 Governor Races

Bloomberg News, in a feature by Stephanie Baker and Priyanjana Bengani, reported that operators of the Russian Matryoshka network (also called Operation Overload) are using automated accounts to inflate view counts on fabricated videos about celebrities and Democratic candidates ahead of the midterms. This entry is built from the full text as carried by The Age, a licensed Bloomberg syndication; Bloomberg’s own page is paywalled and was not read.

Who did the analysis. Bloomberg reviewed 130 videos and doctored newspaper front pages published by the network between September 10 and October 9, most flagged by the anonymous group Antibot4Navalny, which shared archived snapshots of posts with Bloomberg. Graphika and dTeam corroborated and supplemented the narratives. The primary data are Antibot4Navalny’s snapshots; the group’s own publication was not located, so the counts below are Bloomberg’s account of them.

Scope. Per Bloomberg’s review, the network pushed false stories about Democrats in 29 of 36 governor races and 20 of 35 Senate races.

How inflation was detected. Examples cited from Antibot4Navalny’s snapshots: one post got 50,839 views and one retweet within 20 minutes, then under 700 views in the hours after. A Taylor Swift/Billie Eilish post got 75,600 views shortly after upload. A clip about Amy Klobuchar showed 155,300 views by X’s count. Most posting accounts had few followers and had been dormant for years, and reposts and likes were a tiny fraction of views. Clemson’s Darren Linvill: “Views don’t normally drop off a cliff.”

Limits stated. Graphika’s Jean le Roux told Bloomberg the clips are probably not affecting the midterms or how people vote, but serve as a measure of success reported back inside Russia. Graphika and Antibot4Navalny said they had seen no evidence of a decline in Kremlin-linked bot activity since Ukraine’s strikes on Russian data centres began; what fell was Yandex crawler traffic. X suspended 80 per cent of the accounts Bloomberg saw and removed other accounts after being flagged.

Corrections to secondary copies. The 200,000-views-in-the-first-hour line is Bloomberg’s own narration, attributed to Antibot4Navalny’s flagging, not a direct quotation from researchers as the Kyiv Post framed it. The earlier entry 2026-10-17–bloomberg-matryoshka-bot-network-inflates-fake-videos-targeting-democratic-candidates-20-senate-29-governor-races was built from Japan Times and Kyiv Post excerpts; this entry carries the fuller sourcing. Background: 2026-09-11–matryoshka-storm-1679-cnn-branded-celebrity-voice-clone-videos-target-democrats-midterms-ms-now.

Sources & Citations

Tiers Tier 1 court records & gov docs · Tier 2 established outlets · Tier 3 regional & specialty press · Tier 4 opinion or single-source. Methodology →
Cite this entry
The Cascade Ledger. “Bloomberg Investigation (Baker and Bengani) Finds Matryoshka Bots Inflating View Counts on Fake Videos Against Democrats in 20 Senate and 29 Governor Races.” The Capture Cascade Timeline, October 17, 2026. https://capturecascade.org/event/2026-10-17--bloomberg-baker-bengani-matryoshka-bots-inflate-view-counts-of-fake-midterm-videos-antibot4navalny-snapshots/