🔍 Read the full analysis: When AI Agents Bring The Cost Of A Break-In To About $50 on ThorstenMeyerAI.com
Get hardware and tech essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
A Reuters investigation, drawing on records from exposed Cairn control panels and research by NorForge, reported that an AI-agent campaign tried more than 1,200 websites and broke into sites belonging to more than 100 companies. NorForge estimated the campaign’s AI costs at roughly $5,000, or about $50 per successful intrusion, while warning that the estimate carries considerable uncertainty. The amount of data accessed or taken has not been established.
An operator used Cairn, an open-source AI-agent platform, to probe more than 1,200 online shops and break into websites belonging to more than 100 companies, according to records reviewed by security firm NorForge and reported by Reuters on October 19. NorForge estimated the campaign’s AI costs at about $5,000 in total, or roughly $50 per successful intrusion; that is a cost estimate, not a confirmed measure of damage or stolen data.
NorForge researcher Snorre Fagerland found internet-accessible Cairn control panels, some without passwords, while searching for agent dashboards. The firm said it identified more than 20 panels it linked to likely malicious activity. The largest activity cluster, which NorForge labels AIA0001, used Chinese-language task instructions. NorForge counted more than 1,100 attack projects across 14 of its panels, while Reuters reported that the campaign tried more than 1,200 sites from September 25 through mid-October.
The agents worked through online shops in the United States, Canada, Australia and Japan. Reuters identified affected websites associated with companies including farm-equipment maker Deere, appliance maker SharkNinja and engine maker Cummins. The reporting relies in part on the attacker’s own exposed Cairn records; NorForge’s report did not name victims. Reuters said it independently reviewed Cairn activity and contacted companies. Two smaller businesses acknowledged compromises, while some other contacted companies did not respond.
NorForge measured successful projects as taking a little under 23 hours from start to goal, with one lasting more than five days. Its estimate of roughly $5,000 for AI use across the campaign works out to around $50 per successful hack, but the report described the estimate’s uncertainty as considerable. Those figures concern estimated AI costs, not the total cost of the operation, the value of any data or the expense of responding to the intrusions.
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
Why Cheap, Persistent Probing Matters
The reported campaign suggests that AI agents can lower the labor needed to run long, repeated website probes. The estimated $50 per successful intrusion is not a universal price for hacking, nor does it show that every attempted attack succeeds. It does point to a practical concern for defenders: an operator may be able to direct software to keep testing many sites while spending less time supervising each attempt.
The reported targets also show why organizations need to account for support sites and third-party web services, not just their main websites. SharkNinja said the affected site was a standalone product-support site, and Cummins described a potential attack on a website operated by a vendor. Automated activity can also add to the volume of alerts that security teams must review. The source material does not establish that the campaign caused a wider service disruption or that sensitive customer records were taken from the named companies.
For businesses, the case raises questions about how to detect and limit automated probing, protect exposed management panels, and track the systems vendors operate on their behalf. Cairn’s involvement also shows that a general-purpose agent platform can be used outside its stated design goals. That does not make the software itself proof of malicious intent; the security risk described here comes from how an operator used it and from exposed control panels.
Products Worth Considering
Ledger Flex + Folio Crypto Hardware Wallet
Trezor Safe 7 Crypto Hardware Wallet
Trezor Safe 5 Crypto Hardware Wallet with Color Touchscreen, Black Graphite
As an affiliate, we earn on qualifying purchases.
How Researchers Found the Cairn Panels
The investigation followed a separate September report by Gambit Security, which described a different agent-driven campaign. Gambit said that campaign compromised at least 27 companies and that at least 600,000 card records were taken from two of them. Those findings concern a separate operation and should not be read as evidence that the Cairn campaign took the same type or volume of data.
NorForge searched for exposed dashboards used to coordinate Cairn agents after reviewing earlier reporting about AI “harnesses”—programs that let models carry out tasks with some autonomy. The firm found control panels that could be viewed from the public internet, including panels without authentication. Fagerland observed activity through those panels and grouped it into clusters; NorForge published its report on October 19, the same day Reuters published its investigation.
Cairn is described in the source material as a general-purpose, state-based problem-solving engine rather than a dedicated hacking product. Its author, who uses the name le1xia0, told Reuters that Cairn was not designed for hacking and that harmful use was not the author’s intention. The operator’s identity remains unknown; NorForge described the actor as Chinese-speaking, which does not establish nationality or location.
““a little under 23 hours””
— NorForge report
Data Access and Attribution Remain Unclear
Reuters could not establish how much data, if any, was taken in the reported campaign. Deere said it had no indication sensitive information was accessed; SharkNinja said a standalone product-support site was hit and the threat was contained. Cummins said it was investigating a potential cyberattack involving a site run by a third-party vendor. These statements describe company responses, not an independently established account of all access or data exposure.
The operator has not been publicly identified, and the available reporting does not establish who directed the activity or where that person was located. The estimate of about $5,000 in AI costs is also uncertain and does not account for every possible operational expense. Nor is it clear from the source material whether every reported target was successfully compromised or what happened to access after each intrusion.
Company Investigations and Exposure Checks
The affected companies’ investigations and remediation efforts will determine whether further details emerge about the systems reached, any data accessed and the duration of access. Cummins said it was investigating the potential attack on a vendor-operated website; Deere and SharkNinja described steps or containment actions to Reuters. No arrests or charges were reported in the source material.
Security teams may also continue checking for exposed Cairn dashboards and similar agent-control panels. NorForge said new panels appeared as it monitored the activity. Further reporting or company disclosures could clarify the campaign’s scope and consequences, but the information currently available does not resolve whether data was taken from the named companies.
Key Questions
What did the AI agents do?
According to Reuters and NorForge, an operator used Cairn agents to probe online shops. The agents attempted to access more than 1,200 sites and were linked to successful intrusions at websites belonging to more than 100 companies.
Did the attacks cost only $50?
No. NorForge estimated roughly $5,000 in AI costs for the campaign, or about $50 per successful intrusion. The firm said the estimate had considerable uncertainty, and it does not represent the full cost of operating the campaign or responding to it.
Was customer data 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; the source material does not confirm data theft from the other named companies.
Who was behind the campaign?
The operator has not been publicly identified. NorForge described the actor as Chinese-speaking, but that does not establish the person’s identity, nationality or location. No arrest or charge was reported in the source material.
Source: ThorstenMeyerAI.com
