Google's Threat Analysis Group just released its Q4 2025 quarterly bulletin, and the numbers tell a stark story about the scale of coordinated influence operations targeting its platforms. The company terminated more than 18,000 YouTube channels, blocked hundreds of domains from Google News, and shut down dozens of ad accounts linked to state-backed actors across Russia, China, and at least a dozen other countries. The report, published by TAG analyst Billy Leonard, reveals how Russia alone accounted for the majority of takedowns, with massive networks linked to known influence factories like the Internet Research Agency still operating at scale.
Google isn't pulling punches when it comes to state-backed influence operations. The company's Threat Analysis Group published its Q4 2025 bulletin today, documenting a sweeping enforcement action that wiped out more than 18,000 accounts across YouTube, Blogger, and Google Ads - the vast majority tied to coordinated campaigns from Russia and China.
The scale is staggering. According to the official TAG bulletin, Russian actors dominated the takedown lists across all three months, with Google terminating roughly 2,500 YouTube channels in October alone that were linked to a single Russian consulting firm. That network was pushing pro-Russia, anti-NATO content in Russian, targeting audiences with narratives critical of Ukraine and Western allies. By December, Google had axed another 1,256 channels connected to the same operation.
But Russia's influence machinery runs deeper than one consulting firm. The Internet Research Agency - the notorious troll farm sanctioned by the US government - still had active channels on YouTube as recently as December. Google terminated nine IRA-linked channels across October and December, all sharing Russian-language content supportive of the Kremlin and critical of Moldova. The timing wasn't coincidental - Moldova held presidential elections in October, and Russian networks flooded the zone with anti-government narratives.
Google blocked over 150 domains from appearing in Google News and Discover feeds throughout the quarter, most of them tied to Russian operations. In one December action, the company removed 85 domains spreading anti-Ukraine content in English and French. Another batch of 22 domains blocked in October were pumping out pro-Russia narratives in eight languages, including Croatian, Japanese, and Macedonian - a sign of how widely these operations cast their nets.
China's influence efforts, meanwhile, show no signs of slowing. Google's ongoing investigation into PRC-linked coordinated inauthentic behavior resulted in the termination of over 10,000 YouTube channels in Q4 alone. The networks uploaded content in Chinese and English focused on China-US relations and foreign affairs, continuing patterns TAG has tracked for months. In October, 3,715 channels went down. By December, that number hit 6,280 - a sharp escalation that suggests either growing activity or improved detection on Google's end.
The geographic spread of influence operations is notable. Google didn't just target the usual suspects. Indonesia saw massive takedowns, with over 4,000 channels removed across the quarter for coordinated activity supporting Indonesian political interests. Azerbaijan lost more than 800 channels spreading pro-government and anti-Armenia narratives. Pakistan, Bangladesh, Belarus, and even UK-based actors all got caught in TAG's dragnet.
Some operations were surprisingly targeted. In October, Google terminated two YouTube channels and an ad account linked to Israel that were sharing English-language content critical of Canada. Another campaign tied to UK-based actors pushed pro-UK political party content while criticizing the United States. The bulletin doesn't name specific parties or provide attribution beyond country links, but the granular targeting shows how influence ops increasingly focus on niche political wedges.
What's clear from the data: platform manipulation isn't slowing down, it's professionalizing. The Russian consulting firm running thousands of channels wasn't a fly-by-night operation - it sustained coordinated activity across three months despite repeated takedowns. China's networks regenerated faster than Google could remove them, with December's 6,280-channel purge nearly doubling October's haul.
Google's enforcement strategy relies heavily on behavior-based detection rather than content moderation. TAG looks for coordinated inauthentic behavior - accounts that work together to artificially amplify messages, often using fake personas or concealed identities. That's why the bulletin lists "campaigns" rather than individual bad actors. Each bullet point represents a network of accounts working in concert, not isolated violations.
The timing of many removals correlates with real-world events. Moldova's elections drew intense Russian focus, with at least six separate campaigns terminated across October and November targeting Moldovan audiences. Armenia saw multiple Russian-linked networks spreading critical content in Armenian throughout the quarter. These aren't random - they're strategic efforts to shape political outcomes during sensitive periods.
Google's transparency around these takedowns stands in contrast to some competitors. While Meta publishes quarterly adversarial threat reports and X (formerly Twitter) has largely abandoned regular influence operation disclosures under Elon Musk's ownership, Google's TAG bulletins provide granular breakdowns of removal actions without much analysis or strategic context.
The bulletin doesn't address false positives, appeals processes, or how many accounts might have been incorrectly flagged. It also doesn't quantify reach - how many people saw content from these networks before removal, or whether the takedowns happened quickly enough to limit exposure. Those details matter for assessing the real-world impact of influence operations versus the effectiveness of platform responses.
What happens next likely depends on how quickly these networks can reconstitute. If China's operations can bounce back from 6,280 channel removals in a single month, the problem isn't just detection - it's the economics of creating new accounts faster than platforms can remove them. Google's Q1 2026 bulletin will show whether this enforcement wave actually disrupted operations or just forced a temporary reset.
Google's Q4 enforcement numbers expose an uncomfortable reality: state-backed influence operations are scaling faster than platform defenses. Over 18,000 account terminations in three months represents either a massive detection breakthrough or a troubling surge in coordinated manipulation - possibly both. The fact that China's networks grew from 3,715 removals in October to 6,280 in December suggests these operations aren't deterred by takedowns, they're adapting around them. As 2026 elections heat up globally, the real test isn't how many accounts Google can remove, but whether removal actually disrupts the influence or just reshuffles the players.
