It borrowed the ad machinery built to sell you sneakers - and pointed it at the people most likely to do harm online.
Somewhere in a quiet office, an analyst is watching a basketball player get threatened by strangers. Not a stadium of boos - a feed of messages, some idle, a few credible. The analyst's job is to tell the difference, flag what matters, and get the dangerous ones taken down before they find their target. The player never sees most of it. That is the point.
This is Moonshot in 2026. The London-founded company started by tracking ISIS recruiters and now sits behind governments, technology platforms, and - as of 2025 - a professional basketball team, reading the online threat environment in close to real time. It does not run a social network. It does not sell a viral app. It studies the people behind the harm, and intervenes.
"Moonshot doesn't moderate content. It studies the people behind it - then does something about it."
By the mid-2010s, radicalization, harassment and disinformation had moved online and learned to scale. Recruiters used the same targeting tools as marketers. Conspiracy theories travelled faster than corrections. Most institutions answered with the blunt instrument of deletion: take the post down, suspend the account, repeat forever. Useful, and roughly as durable as bailing a boat with a fork.
The founders - both extremism researchers - had spent years interviewing the people on the other side of that screen. They had a hunch the establishment found uncomfortable: you cannot moderate your way out of human behaviour. You have to reach the human.
"You cannot delete a belief. You can, occasionally, interrupt one."
The numbers above are Moonshot's own. They are the kind of statistics that ruin a dinner party and start a company.
Vidhya Ramalingam and Ross Frenett founded the company in 2015 and called it Moonshot - a wink at the kind of goal that sounds impossible until someone insists on trying. Ramalingam had led global responses to extremism and would later be named an Obama Foundation Leader. Frenett had directed a worldwide network of former extremists, defectors and survivors - people who had walked out of the very ideologies the company wanted to disrupt.
Their bet combined two things that rarely share a room: the cold precision of data science, and the patience of a counsellor. Measure the online threat with engineering rigour. Reach the person at risk with something closer to empathy. Then prove, with baselines and evaluations, whether any of it actually worked.
"The name riffs on the impossible goal. The method is stubbornly practical."
Co-Founder & CEO. Recognised internationally for leading responses to extremism; Obama Foundation Leader: Europe (2020).
Co-Founder. Former director of a global network of ex-extremists and survivors; took the company from conflict zones to government briefing rooms.
Two extremism researchers turn field experience into a company built to counter violent extremism online.
Targeted ads meet at-risk searchers - steering them toward counter-narratives instead of recruiters.
The method is adapted to new threats; the company expands into the wider field of online harms.
Vidhya Ramalingam joins the inaugural Leaders: Europe cohort.
Beringea and Mercia back the company to scale its online-threat technology.
The WNBA's Chicago Sky becomes the first team to deploy Moonshot to protect players from online abuse.
The idea that made Moonshot famous is almost embarrassingly simple. When someone types a phrase that signals they are sliding toward extremist content, the Redirect Method answers - not with a lecture, but with an ad. The ad leads somewhere useful: a counter-narrative, a former extremist's testimony, a helpline, a moment of doubt. Built with Google's Jigsaw in 2016, it was first aimed at ISIS and later turned on white-supremacist and disinformation networks.
Around that core sits the rest of the toolkit: OSINT-driven threat monitoring, behavioural analytics and machine learning to sort noise from credible risk, dedicated teams to get dangerous content removed, and evidence-based interventions for both victims and would-be perpetrators. The unglamorous discipline underneath it all is measurement - baselines, evaluations, and the willingness to ask whether an approach worked.
Targeted ads steer searchers of harmful content toward safer alternatives and trusted voices.
OSINT monitoring that detects, assesses and prioritises online risk across platforms.
Teams that flag concerning content, judge credible threats and help get harm removed.
De-escalation, perpetrator engagement and crisis support for people at risk.
"Answer a dangerous search with a better question. Then measure whether the question landed."
The customer list is the argument. Governments in the UK, Canada and Japan have paid Moonshot to counter extremism and disinformation. Technology companies - Google, Meta and Spotify among them - have funded or partnered on the work. And in 2025 the field widened again when the WNBA's Chicago Sky became the first professional team to hire Moonshot to shield its players, after a season in which the abuse aimed at women athletes stopped being background noise.
Bars scaled for display. Abuse figures are percentages; funding and headcount are absolute values shown for context.
"The same firm that once tracked recruiters now helps shield point guards from trolls."
UK, Canada, Japan and others - paid programmes against extremism and disinformation.
Google & Jigsaw, Meta, Spotify - funding and trust-and-safety collaboration.
U.S. Olympic & Paralympic teams, NCAA programmes, and the WNBA's Chicago Sky.
A company that reads the worst of the internet for a living has to be careful about how it does the reading. Moonshot leans on ethical data sourcing, GDPR and data-privacy compliance, and a "safety by design" posture - the reassuring corporate language for a genuinely thorny question: how do you study harmful people without becoming a surveillance shop. The answer it has chosen is to point the work at protection - of victims first, and of the people who have not yet crossed a line.
The scope kept growing because the harm did. What began as counter-extremism now spans disinformation, gender-based violence, human trafficking, and child sexual exploitation. Different threats, one through-line: find the people, understand the drivers, intervene early.
"Protect the victim first. Reach the person on the brink second. Delete as a last resort."
Back in that quiet office, the player keeps playing.
The threats still come - the internet hasn't reformed. But now someone is reading them, sorting them, and pulling the dangerous ones down before they land. That is the whole bet, ten years on: harm at scale, answered by people who insist on understanding it. The moonshot was never to delete the dark. It was to get there first.