The upcoming midterms are widely seen as a referendum on Donald Trump’s presidency, but its defining issue to date is notably MIA. “Campaign ads and debates are mostly avoiding the Russia investigation,” Politico reports, “in favor of other issues important to voters…like the economy, health care and taxes.” One study of political ads over a four-week period through mid-October found that 0.1 percent of ads aired in congressional races mentioned Russia; there were zero mentions of Russia in ads for Senate races.

On one level, it is unsurprising that the election has been focused on issues that impact voters’ lives, rather than the byzantine bureaucratic drama that has consumed Washington and elite media since Trump’s election. But after months of fearmongering about a sweeping Russian interference effort and a compromised, complicit president, perhaps we are also seeing the penny start to drop: Russiagate, for all its hype, has not gone as advertised.

Take the supposed Russian threat to the midterms. For months, intelligence officials and prominent media outlets have bombarded us with warnings about “a pervasive messaging campaign by Russia to try to weaken and divide the United States” (Director of National Intelligence Dan Coats), a threat so dire that we might as well dub the vote the “The Moscow Midterms” (FiveThirtyEight) and acknowledge that “we’re defenseless against Russian sabotage in the midterm elections,” (Washington Post columnist Jennifer Rubin). The New York Times informed readers in July that Coats had likened “the persistent danger of Russian cyberattacks today…to the warnings the United States had of stepped-up terror threats ahead of the Sept. 11, 2001, attacks.” “The warning lights are blinking red again,” he said.

While it is always possible that new evidence of interference will emerge, so far, the Russian danger has had an underwhelming denouement. Russia’s alleged midterm sabotage to date has been disclosed in a newly unsealed criminal complaint against an employee of the Internet Research Agency (IRA), the Russian troll farm previously indicted for using fake accounts to spread divisive content on social media. The defendant, Elena Khusyaynova, is not even directly accused of online manipulation. Instead, she is singled out for being the chief accountant for “Project Lakhta,” an IRA initiative that targeted audiences in Russia and around the world, including the United States. The only actions directly ascribed to Khusyaynova concern her “meticulous record-keeping and management” of IRA funds.

As with the initial indictment of 13 IRA employees in February, prosecutors accuse IRA trolls of using social media “to sow discord in the U.S. political system and to undermine faith in our democratic institutions,” including in the upcoming midterms. But reading the fine print, it is difficult to see how that purported aim can be taken seriously. In the first six months of 2018, Khusyaynova submitted expenditures of $60,000 for advertisements on Facebook and $6,000 on Instagram. The only advertising-related activity that gets detailed in the complaint is an alleged IRA employee’s offering to give organizers of an anti-Trump protest $80 for a Facebook ad. (It’s unclear if the proposal was accepted). The IRA’s alleged social-media accounts impersonated both liberal and conservative personas. The complaint shows six images that were posted by Russian trolls to Facebook; the most impactful of the bunch appears to be a thinly disguised anti-Muslim ad that attracted 104 comments.

Some of this content is overtly racist and bigoted; other posts are banal. All are so juvenile or inconsequential that it is difficult to see how they could have vastly greater influence than the millions of other pieces of political clickbait littering the Internet. The IRA’s social-media imprint seems to have as much impact now as it did during the 2016 election. Back then, the IRA spent a reported on $100,000 on Facebook ads, with most of those ads having nothing to do with the election, and more than half of that total spent after the election.

According to Facebook general counsel Colin Stretch, posts generated by suspected Russian accounts between 2015 and 2017 represented “a tiny fraction of the overall [News Feed] content on Facebook.… about four-thousandths of one percent (0.004%) of content in News Feed, or approximately 1 out of 23,000 pieces of content.” The widely cited figure that “material generated by the Kremlin had reached a hundred and twenty-six million American Facebook users,” (The New Yorker) is in fact a creative take on Facebook’s own speculation. “Our best estimate,” Stretch testified to Congress in October 2017, “is that approximately 126 million people may have been served one of these [IRA] stories at some time during the two year period.” So the 126 million figure is an “estimate” of how many people “may have been served” one piece of IRA content—most unrelated to the 2016 election—in their Facebook feeds over two years. Over on Twitter, a new analysis by the Atlantic Council’s DFRLab finds that “Russia’s troll operation primarily targeted Russian speakers,” posting “significantly more in Russian than in English.”

Unsurprisingly, none of this has slowed the histrionics. The unsealed complaint against the Russian troll-farm accountant prompted Mother Jones to declare that “Russia is Now Attacking the Midterm Elections,” and NPR to warn audiences of “How Russia Runs Its Disinformation Effort Against The 2018 Midterms.” The ranking Democrat on the House Intelligence Committee, Representative Adam Schiff, called the new activity a sign that “Russian disinformation efforts are ongoing and sophisticated, and they are intent on dividing us and weakening our society and political system.” But given what we actually know about Russian disinformation, its most significant impact appears to be as fodder for ongoing efforts intent on convincing Americans that unsophisticated social-media trolling could somehow divide and weaken their society.

The alarm about Russian social-media trolls has been complemented by warnings that Russia could penetrate state voting systems, as it allegedly did to 21 states before the 2016 elections. But as journalist Gareth Porter notes, the 21-states figure was based on a preliminary assessment that itself was based on yet another guess. As senior DHS officials explained to Congress, “21 states were potentially targeted by Russian government cyber actors,” and even that assessment was “not a definitive source in identifying individual activity attributed to Russian government cyber actors.” Several states subsequently disputed that they were “targeted” at all, and another senior DHS official described “the majority of the activity” as “simple scanning…a regular activity across the Web. I would not characterize that as an attack.”

Whatever it was, the only public evidence for claims that Russian government cyber actors targeted US voting systems is a leaked National Security Agency document reported by The Intercept in June 2017. It accused Russian military intelligence of impersonating a voting software company and sending malware-laden e-mails—a widespread hacking tactic known as spear-fishing—to more than 100 state and local jurisdictions. The Intercept acknowledged that the NSA document “does not show the underlying ‘raw’ intelligence,”—i.e., the evidence—”on which the analysis is based.” But The Intercept failed to note that the document doesn’t even accuse the Russian government with certainty. The NSA’s “analysis” that the spear-phishing came from Russian military intelligence is not an attribution based on “Confirmed Information,” the document disclaims, but instead on “Analysis Judgement” (sic) and the even less convincing “Contextual Information.” The hackers were also not so inconspicuous: Their fake VR Systems e-mails were sent from a Gmail account: “vrelections@gmail.com.”

The underwhelming nature of Russia’s alleged cyber-operations also dovetails with the heretofore-empty quest to uncover whether Trump and members of his circle are criminal accomplices. As noted in a sweeping, 10,000-word New York Times piece aimed at “Unraveling the Russia Story So Far,” “no public evidence has emerged showing that [Trump’s] campaign conspired with Russia in the election interference or accepted Russian money.”

In the place of evidence, the oddest theories have somehow become fodder for enlightened discussion. This past summer, the possibility that Trump has been a Russian intelligence asset since 1987 was the subject of an extensive piece by New York magazine’s Jonathan Chait, complete with an obligatory conspiracy map and the even more obligatory coverage on MSNBC. A recent lead story in The New Yorker investigated the rumor that the Trump campaign used a computer server to secretly communicate with Russia before the 2016 election. The FBI has previously looked into the claim and dismissed it, but there are others on the case: The New Yorker’s story is based on the suspicions of a group of anonymous computer scientists who deem themselves to be “self-appointed guardians of the Internet.”

Their theory is that DNS traffic between a Trump marketing server and Russia’s Alfa Bank is a sign of “a covert communication channel.” The piece contains caveats familiar to the Russiagate canon: The available data prevent researchers “from going beyond speculation,” and no one “could be certain of what Alfa Bank and the Trump Organization were doing.” What is perhaps most remarkable about the story is its acknowledgment that the Trump Organization did not even control its own server. Instead, the server that it would have used to secretly chat with Russian conspirators was managed by Pennsylvania marketing firm Listrak, “which mostly helped deliver mass-marketing emails: blasts of messages advertising spa treatments, Las Vegas weekends, and other enticements.”

The outlandish speculation is a part of a wider trend that has fueled the Russiagate story from the start. This fixation has also meant that vital, even existential issues, are thrown to the side, including major developments that contradict its underlying collusion narrative. Over the weekend, Trump announced that the United States will pull out of the 1987 Intermediate-range Nuclear Forces treaty, a bedrock of international security. The move prompted outrage from Russia and warnings of “the most severe crisis in nuclear arms control since the 1980s,” (Malcolm Chalmers of the Royal United Services Institute), including the threat of a new nuclear arms race. The move is widely credited to National Security Adviser John Bolton, who, as I warned in The Nation in April, sees the Russiagate moment as “a not-to-be-missed opportunity” to advance a hawkish neoconservative agenda, including the dismantling of arms-control treaties that he began under George W. Bush.

Despite this dangerous move, none of the prominent voices who have accused Trump of being soft on Russia, speculated whether he is a Kremlin operative, is vulnerable to Russian blackmail, or is even taking orders directly from Putin, have stepped forward to revisit their collusion theory. Worse, some are doubling down. Decrying Trump’s failure to confront the Kremlin in the wake of the Russian troll farm accountant’s indictment, Representative Schiff told MSNBC that Putin must think to himself, “This weak U.S. President will never confront me—he doesn’t have the guts to confront me.”

As diplomats and arms-control experts confront Trump’s abrogation of a vital nuclear treaty, Schiff and other prominent Russiagate exponents are notably silent. The reasons seem clear: Ever since the 2016 election, the figureheads of Trump’s political and media opposition have invested in a supposition that Trump is in cahoots with Russia and encouraged him to be confrontational as a means of disproving it. They have de-incentivized and disempowered themselves to stand up to Trump when he confronts Russia in one of the most reckless ways that he could. For all the dire warnings about Russian trolls and hackers over the past two years, it is those sounding the alarms who have fueled a much worse threat.