🔍 Read the full analysis: AI Agents And The New Economics Of Hacking Company Websites on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
A Chinese-speaking operator used the open-source Cairn platform to coordinate AI agents that probed more than 1,200 websites from late September into October. Reuters, citing the attacker’s records, reported break-ins at more than 100 companies, but the amount of data accessed or taken is unknown.
A Chinese-speaking operator used the open-source Cairn AI-agent platform to probe online shops from September 25 into mid-October, with Reuters reporting break-ins at websites belonging to more than 100 companies, including Deere, SharkNinja and Cummins. NorForge and Reuters say the agents tried more than 1,200 sites; how much information, if any, was accessed or taken remains unclear.
NorForge researcher Snorre Fagerland found exposed Cairn control panels online and monitored their activity. NorForge linked more than 20 panels to likely malicious use and grouped their activity into three clusters. The largest, designated AIA0001, used Chinese-language task instructions. Across 14 panels in that cluster, NorForge counted more than 1,100 attack projects, aimed mainly at targets in the United States, Japan and Australia. Reuters reported that the broader campaign also targeted sites in Canada.
NorForge describes Cairn as a general-purpose, state-based problem-solving platform: agents record findings and planned next steps on a shared board, allowing other agents to continue the work. The firm says the campaign fed shop websites into an automated process before agents investigated them. For successful AIA0001 projects, NorForge measured an average duration of a little under 23 hours; one lasted more than five days. Those figures describe the observed projects, not a general benchmark for AI hacking.
Reuters identified companies by reviewing the operator’s Cairn activity and contacting potential victims. Deere told Reuters it took “necessary remediation actions” and had “no indication that sensitive information was accessed.” SharkNinja said a standalone product-support site was affected and the threat was contained. Cummins said it was investigating a potential attack on a website run by a third-party vendor. Reuters could not establish whether data was taken or how much.
One operator, a swarm of agents
The campaign in numbers
What the researcher and Reuters report about the busiest group NorForge found. Counts come from the attackers’ own exposed panels; cost and success figures are estimates.
The price tag estimates
NorForge / Fagerland, “uncertainty considerable”
- ~$5 per task on average, including tasks that fail early
- $100+ for long projects with more than 1,000 intents and facts
- ~$5,000 in AI costs for the campaign, about $50 per successful hack
- ~10% success rate, NorForge’s “reasonable assumption”
How the swarm runs
Simplified, no attack details
- Shop websites are ranked by traffic and checked automatically in a browser
- Each target becomes a Cairn project: a starting point and a goal
- Agents post facts and next steps on a shared board and build on each other
- Goal: a hidden way back in to shop and payment systems, for later
- 22 SepGambit reportGambitCairn found among three AI harnesses on a criminal’s staging server (separate campaign)
- 25 SepCampaign startsReutersA Chinese-speaking operator turns agents on shops in the US, Canada, Australia and Japan
- ≈ 2 OctOpen panels foundReutersFagerland starts finding Cairn panels in malicious use, about two dozen
- mid-Oct1,200+ sites triedReutersBreaks into websites of 100+ companies, incl. Deere, SharkNinja, Cummins
- 19 OctReport and investigationNorForge · ReutersNorForge publishes “Inside Cairn”; Reuters publishes its investigation
Sources: NorForge, Inside Cairn (19 Oct 2026); Reuters (19 Oct 2026); Gambit Security (22 Sep 2026). Numbers as published, same rounding as the article.
Powered by Thorsten Meyer AI
Lower Costs, Broader Exposure
The reported campaign points to a change in the economics of persistent intrusion attempts: one operator can coordinate many automated agents to keep testing sites over extended periods. NorForge estimates the campaign used roughly $5,000 in AI services, or about $50 per successful intrusion. The firm cautions that the estimate carries “considerable” uncertainty; it is not a confirmed accounting of the operator’s costs.
For organizations, the reported victims include not only major manufacturers but also the websites and services around them. Product-support pages and third-party vendor sites can be connected to a company without being its main customer-facing platform. The company statements describe different scopes of impact, and the available reporting does not establish that every reported break-in exposed sensitive data.
Automation can also increase the volume of activity defenders must review. The concern is not that AI agents make every intrusion successful, but that they may let an operator sustain more attempts with less direct attention. The campaign’s actual financial returns, if any, have not been established.
Products Worth Considering
DJI Osmo Mobile 7P Phone Gimbal with Mic Mini Transmitter, Infinity Black
LG UltraGear 32GX850A-B 32-inch 4K UHD Glossy OLED Gaming Monitor
DJI Mic 3 Wireless Microphone System (2 Transmitters, 1 Receiver, Charging Case)
As an affiliate, we earn on qualifying purchases.
How Researchers Traced Cairn
The investigation followed a separate report published in September by Gambit Security. That firm said it recovered a criminal staging server containing three tools that let AI models act autonomously: Hermes, Strix and Cairn. NorForge then searched for Cairn dashboards exposed to the internet. Its report says most panels require authorization, but researchers found some without passwords and began identifying panels associated with suspicious activity around October 2.
NorForge published its findings on October 19, 2026, the same day Reuters reported its investigation. NorForge’s report did not name victims; Reuters said it identified companies by independently reviewing the attacker’s Cairn activity and contacting them. The two accounts provide evidence about different aspects of the operation: NorForge analyzed exposed panels and project records, while Reuters reported victim identities and company responses.
The use of Cairn does not by itself show that the platform was designed for cybercrime. Its author, who uses the name le1xia0, told Reuters it “was not designed for hacking attacks” and said harmful use was not the author’s intention. The reported campaign concerns how an operator used the tool, rather than proof that Cairn’s intended purpose was malicious.
““a little under 23 hours””
— NorForge report
Data Access Still Unverified
The available reporting does not establish what data, if any, was accessed or removed from the affected company websites. Deere said it had no indication of sensitive information being accessed; that statement does not establish what happened at other victims. SharkNinja described a standalone support site as affected, while Cummins said it was investigating a potential attack involving a third-party vendor website.
The operator’s identity is also unknown. The source material describes the actor as Chinese-speaking, but does not establish who operated the campaign or where that person was located. Reuters reported no arrest or charge. NorForge’s financial estimate is uncertain, and the record does not establish whether the intrusions produced revenue or persistent access in every case.
Victim Reviews and Investigation
The companies’ statements indicate different responses: Deere said it took remediation steps, SharkNinja said it contained the threat, and Cummins said its investigation was ongoing. The extent of any data exposure, whether access persisted after discovery, and whether other organizations were affected remain subjects for investigation. No later milestone or enforcement action is established in the available reporting.
For readers tracking the story, the clearest next developments would be further findings from the affected companies, additional technical reporting on the exposed Cairn panels, or information from investigators about the operator. Until then, the reported breach count and cost estimate should be read alongside their stated limits: Reuters tied the company count to the attacker’s records, and NorForge described its cost calculation as uncertain.
Key Questions
What happened in the reported campaign?
A Chinese-speaking operator used Cairn to coordinate AI agents probing online shops. Reuters reported that the operation tried more than 1,200 websites and broke into sites belonging to more than 100 companies.
Which companies were named?
Reuters named Deere, SharkNinja and Cummins among the companies whose websites were affected. The companies described different sites and responses; the reporting does not establish the same impact for each.
Was sensitive information stolen?
That has not been established. Reuters could not determine how much data, if any, was taken. Deere said it had no indication that sensitive information was accessed, while other company investigations or statements address different parts of the incidents.
Was Cairn built as a hacking tool?
The source material describes Cairn as a general-purpose, open-source agent platform. Its author told Reuters it was not designed for hacking attacks and said harmful use was not the author’s intention.
How much did the campaign cost?
NorForge estimated about $5,000 in AI costs, or roughly $50 per successful intrusion, but said the estimate had considerable uncertainty. It is an estimate, not a confirmed accounting of the operator’s spending.
Source: ThorstenMeyerAI.com
