Yesterday
On June 7, 2026, Armenian voters went to the polls. Prime Minister Nikol Pashinyan’s Civil Contract party won 49.8% of votes — enough for a parliamentary majority. At 2:49 AM on June 8, Pashinyan declared victory, calling it a “historic win” and a vote for “regional prosperity and cooperation.”
The pro-Russian Strong Armenia party came second with about 23%. Former president Robert Kocharyan’s Armenia Alliance took around 10%.
Pashinyan won. The disinformation campaign targeting him ran for three months, produced 343 fake videos, impersonated major Western news outlets, and was described by researchers as the most extensive operation of its kind in recent years — second only to what was deployed against Moldova during its 2025 election.
The campaign failed to flip the election. That is not the same as saying it failed.
Why Armenia
Armenia has been moving. After decades inside Russia’s strategic orbit — a CSTO member, host to Russian military bases, deep economic ties — Pashinyan’s government began, quietly and then less quietly, turning West. Talks with the EU. Zelensky’s visit to Yerevan. Formal applications toward European structures. Suspension of CSTO participation.
From Moscow’s perspective, this is not an internal Armenian matter. It is a defection. The former Soviet periphery is the one resource Russia cannot afford to exhaust — each country that successfully pivots West is both a loss and a signal to others. If Armenia completes the turn, it demonstrates that the turn is possible.
The disinformation campaign that ran against Pashinyan is not best understood as an attempt to manipulate one election. It is best understood as the cost Russia imposes on the attempt.
The Operation
The campaign, identified and tracked by researchers including the Antibot4Navalny collective, Clemson University’s Media Forensics Lab, Germany’s BND, and France’s Viginum, involved two overlapping operations.
Matryoshka is a pro-Kremlin disinformation network that has been running since at least 2024, named for its nested structure: fake sources cite fake sources, building an apparent chain of corroboration that collapses if you pull any single thread. The Armenia campaign represented what researchers described as an “expanded” use of artificial intelligence within Matryoshka — AI-generated video content at a scale that would not have been feasible with human production teams a few years ago.
Storm-1516 is Russia’s most impactful information operation. Between April 2025 and April 2026, Armenia was targeted by Storm-1516 more times than any other country. The operation specializes in manufactured credibility: creating fake regional media outlets, falsifying content from reputable international publications, and routing disinformation through third countries to obscure the origin and make the content appear to arise from the local information environment rather than from an external actor.
The combination is precise. Matryoshka provides the nested structure — each fake appears to cite another source. Storm-1516 provides the credibility laundering — the sources appear to be independent local outlets, not Kremlin contractors. AI provides the volume — 343 videos across three months is a publication rate that a staffed disinformation operation would have found expensive; a generation pipeline finds it routine.
What the Fake Broadcasts Said
The content of the campaign was not random. It targeted specific vulnerabilities in the Armenian political moment.
A fake Euronews broadcast accused Pashinyan of misconduct. The visual design replicated Euronews’s actual broadcast format; the content was fabricated. A viewer encountering the clip without context — shared on Telegram, posted on X, forwarded in a group chat — had no immediate signal that the outlet was counterfeit. Euronews’s brand is the message: a credible European news organization has found evidence of wrongdoing.
Fabricated videos claimed a “secret deal” between Pashinyan and French President Macron. The specific content of the deal was unspecified — the accusation structure requires only that something hidden be implied. Secret deals between Western leaders and a pivoting Armenian PM activate existing suspicion about what Westernization really means: foreign dependency, hidden obligations, national sovereignty quietly traded for EU proximity.
A fake claim spread on May 11 that Pashinyan’s own press secretary had confirmed the presence of NATO instructors in Armenia. The sourcing detail — the press secretary as source — is a specific credibility tactic. Not “pro-Western sources say” but “his own spokesperson confirmed.” The target is not people who already support Russia; it is Armenians who support Pashinyan but hold anxieties about what his Western alignment actually commits them to.
False allegations that Pashinyan misused €9.5 million in EAEU (Eurasian Economic Union) digitalisation funds for campaign financing were circulated through accounts that presented themselves as independent Armenian civil society observers. The figure is precise enough to be checkable. The checking takes time. The allegation spreads faster.
The accumulated effect of 343 such videos is not any single claim. It is the cumulative impression that something is wrong — that something is being hidden, that the EU relationship is transactional on the Western side, that trusting Pashinyan means accepting risks that aren’t being disclosed to you.
The Laundering Architecture
Third-country laundering deserves closer attention than it typically receives.
A disinformation video that clearly originates from a Russian state operation is received differently than one that appears to originate from a Bulgarian blogger, a Georgian civil society account, or a neutral-seeming Armenian news aggregator. Storm-1516’s documented method involves routing content through proxies in countries that are not parties to the underlying dispute — the content appears to emerge from the broader information environment rather than from Moscow.
This is not simply about hiding the origin. It is about changing the epistemics of the claim. A viewer who encounters “Russian state media says Pashinyan made a secret deal with Macron” can apply an appropriate skepticism: the source has an obvious stake in this narrative. A viewer who encounters the same claim attributed to a Bulgarian investigation site that appears to have covered Balkan politics for three years arrives at the claim with a different evaluative frame. The third-country structure manufactures the appearance of independent corroboration.
When multiple fake outlets in different countries are all amplifying the same narrative, with different bylines, different visual designs, different apparent editorial identities, it begins to look like a story that’s getting traction. The saturation creates the false appearance of independent convergence.
Accountability
On May 11, the United Kingdom sanctioned the Social Design Agency (SDA) and 49 of its employees, along with the group ANO Dialog and its leadership.
The SDA had been tasked by the Kremlin to deliver interference operations “designed to undermine democracy and weaken support for Ukraine.” Their staff — writers, translators, video producers — operated what UK officials described as a system that mimicked Western news outlets to spread disinformation. In the Armenian context, London accused the SDA of attempting to install pro-Moscow figures ahead of the June 7 elections. ANO Dialog was alleged to have coordinated with Russian intelligence on Armenia’s domestic politics.
The sanctions are meaningful as documentation. They establish, on the public record of a Western government, that specific named individuals at specific named organizations were employed to do this work. That is different from a research report attributing the operation to “pro-Kremlin actors.” It is a government formally identifying the people.
What sanctions do not do is stop the operation. The videos had been circulating for two months before the sanctions were announced. The election ran on schedule. The sanctioned individuals continued to operate from Russia.
What the Third Means
Pashinyan won. The pro-Russian Strong Armenia party got about 23%. The Armenia Alliance took around 10%. Combined, roughly a third of Armenian voters cast ballots for parties that represented, in varying degrees, a reversal of Pashinyan’s Western turn.
The disinformation campaign cannot be credited for this vote share. Armenian domestic politics existed before Storm-1516. There are genuine constituencies in Armenia with genuine reasons to be skeptical of rapid Westernization — concerns about what EU alignment costs, uncertainty about what replacing Russian security guarantees with Western proximity actually provides, memory of how quickly Western support materialized or didn’t materialize during the 2020 war with Azerbaijan.
But the campaign was not designed to manufacture a pro-Russian majority. It was designed to work on the margin, and on the long game. Every percentage point of opposition that remains after a Western-aligned government wins is leverage: for post-election instability claims, for “democratic legitimacy” questions, for the ongoing narrative that the Armenian public is divided about its direction, that Pashinyan does not speak for all Armenians, that the pivot is contested.
The fabricated videos about EAEU fund misuse do not need to produce a court case to serve their function. They need to be circulating in the Armenian information environment when someone who is slightly skeptical of Pashinyan encounters them. The skepticism was already there. The disinformation is designed to confirm it, solidify it, and make it feel like it’s backed by evidence — fake evidence that operates exactly like real evidence in the absence of immediate verification.
Influence Tactics Breakdown
Source Identity Manufacture. Fake Euronews broadcasts, fake regional media outlets, content routed through third-country proxies — each layer is designed to make the claim appear to originate from a credible independent source. The manipulation is not in the content alone; it is in the constructed credibility of the apparent source.
Third-Country Laundering. Routing disinformation through accounts and outlets in uninvolved countries manufactures the appearance of independent corroboration. When the same narrative appears in sources from multiple countries, it creates false evidence of convergent discovery — each outlet appears to have arrived at the story independently.
Precision Vulnerability Targeting. The specific narratives — NATO instructors, secret deals with Macron, EAEU fund misuse — are not random. They are calibrated to Armenian political anxieties about what Westernization actually means for sovereignty, security, and institutional integrity. The disinformation doesn’t push claims that have no purchase; it amplifies pre-existing doubts with manufactured evidence.
AI Volume Legitimacy. 343 videos across three months is not a human-scale production rate. It is a pipeline. The volume creates the appearance of sustained investigative attention — as if multiple outlets, across multiple weeks, are all independently finding evidence of the same pattern of behavior. That appearance is fabricated, but it functions as a credibility signal.
Post-Election Utility. An operation targeting an election is not a failure if the election is lost by the operation’s preferred outcome. The vote share for opposition parties becomes raw material for post-election delegitimization: claims of fraud, evidence of a divided public, ongoing narrative pressure that the winner lacks a true mandate. The disinformation seeds doubt; the vote share provides the number.
The Pattern
Moldova’s 2025 election saw the most extensive Russian disinformation campaign on record up to that point. Armenia 2026 is the second most extensive on record. The common variable is not the country — it is the direction.
Both Moldova and Armenia, under their current governments, are formally moving toward European structures. Both have significant Russian-speaking populations with historical and cultural ties to Russia. Both sit in the space between Russian and European spheres of influence.
The campaigns against them are not revenge. They are operations conducted at specific strategic inflection points — the elections that could, if lost, make the turn permanent or, if won with a weakened mandate, create conditions for reversal.
The countries that have not seen these campaigns are the ones that haven’t tried to leave.
That is what the “pivot tax” is. Not a one-time cost. A structural reality: any country that tries to change its geopolitical alignment while remaining within reach of the apparatus will face a campaign designed to make the change as costly as possible — in public trust, in institutional stability, in the quality of its information environment.
Pashinyan won. The apparatus will not stop.
This article is part of Decipon’s Manipulation Breakdowns series, examining specific influence operations through the Influence Tactics Protocol.
Sources:
- Pro-Kremlin actors launch large scale disinformation campaign targeting Armenia’s elections — Euronews
- U.K. Sanctions Russian ‘Disinformation’ Outfit Over Plot to Sway Armenian Elections — The Moscow Times
- Russia Escalates Early Election Interference in Armenia — FDD
- Armenia’s Pashinyan Seals Election Win — US News
- Countering Foreign Influence Operations in Armenia — IPHR