AI Agents Hacked 100+ Companies for About $50 Each

AI agents hacked 100+ companies for about $50 per break-in, while one hacker barely lifted a finger. Here is what happened, and what it means for every website owner.

What Happened: One Hacker, 100+ Break-Ins

On October 19, 2026, Reuters reported a striking fact. A single Chinese-speaking hacker broke into websites belonging to more than 100 companies, including farm equipment maker Deere, home appliance company SharkNinja, and engine manufacturer Cummins. The hacker did almost none of the work himself. He handed the job to artificial intelligence and let AI agents do the hacking for him.

The discovery came from Snorre Fagerland, principal threat researcher at the cybersecurity company NorForge. Fagerland was scanning the internet for accidentally exposed installations of Cairn, an open-source AI problem-solving platform popular with hackers. Beginning around October 2, he found about two dozen Cairn control panels left open on the internet and being used for malicious activity. Because those panels were exposed, he could quietly watch the hackers' AI tools at work.

The campaign began on September 25, 2026, and targeted e-commerce websites in the United States, Canada, Australia, and Japan. By late last week, the agents had attempted to break into more than 1,200 sites and had succeeded against more than 100 of them. Fagerland estimated the whole campaign cost roughly $5,000 in AI tokens, or about $50 per successful hack. Each successful break-in typically took around 23 hours of probing. Fagerland's report did not name the victims, but Reuters independently reviewed the hacker's Cairn activity and identified Deere, SharkNinja, and Cummins among them.

How AI Agents Hacked 100+ Company Websites

What Is an AI Agent?

An AI agent is software that can pursue a goal on its own, step by step, instead of waiting for a human to tell it what to do next. You give it a task, and it plans, tries things, checks the results, and adjusts. Here, the "task" was breaking into websites. The agents searched for weaknesses, tested them, and kept going until they found a way in, without a human watching over them.

The campaign ran on Cairn, which coordinates groups of agents working on the same problem. According to Reuters, the operation appeared to use a mixture of China's DeepSeek model and Anthropic's Claude Code. The agents shared a task board, which is what let Fagerland observe the campaign in real time.

Step 1: A Tiny Set of Instructions

The hacker's instructions were astonishingly short: fewer than 100 Chinese characters sketching out his preferred intrusion tactics. In plain terms, the orders boiled down to this: find a way to run code on each site, then plant an uploader web shell, a small hidden program that lets an attacker control a server through a web browser. That short instruction was all the agents needed.

Step 2: Patient, Tireless Probing

Once set loose, the agents spent hours methodically probing each target. In one case Fagerland described, the agents found no weakness in a big company's public-facing application. Instead of giving up, they dug through digital records, found associated infrastructure, located a public-facing control panel, and found a subtle flaw that let them trick the server into running malicious commands. Fagerland, who has worked in cybersecurity for nearly three decades, said: "It's not straightforward to come up with solutions like that. It's not script kiddie stuff."

Who Was Hit, and What Was Taken?

Deere said it had taken remediation actions and had no indication sensitive information was accessed. SharkNinja said only a standalone product-support site was hit and the threat was contained. Cummins said it was investigating a potential attack on a website run by a third-party vendor. Two smaller businesses also acknowledged being compromised.

One important fact remains unknown: Reuters could not determine how much data, if any, was actually taken. A break-in does not automatically mean a major data theft, but it does mean the attacker had the opportunity. Anyone who shopped on an affected site should stay alert for unusual account activity.

The Chinese-speaking hacker was not the only one Fagerland observed. One of the other hackers he watched stole a trove of credit card data. Another tried to get AI agents to rob a cryptocurrency exchange.

Why This Changes Cybersecurity

This campaign matters because it changes the economics of hacking. Until now, breaking into 100 companies required a team of skilled hackers, weeks of work, and serious money. This time, one person spent about $5,000 on AI tokens and let software do the rest. The agents do not sleep, do not get bored, and do not charge by the hour.

It also changes who can hack. The operator did not need to personally find each weakness or write each exploit. He needed a short instruction and access to AI tools that are commercially available. Fagerland had previously assumed hacking agents would still need some human supervision. This campaign proved otherwise. "It's a new world now," he said. "The development in this field is astounding."

AI agents are becoming a standard tool in the cybercriminal underground, not a rare experiment. For more on this fast-moving area, see our more cybersecurity news and more AI news pages.

What Website Owners Should Do Now

If you run a website, especially an online store, this story is a direct warning. AI agents are now probing sites like yours around the clock:

• Patch everything, fast. Keep your website software, plugins, themes, and server fully updated. Agents hunt for known flaws first.
• Take admin panels off the public internet. Never leave control panels, dashboards, or AI tool interfaces exposed without a login.
• Turn on multi-factor authentication (MFA). Require a second verification step for every admin and hosting account.
• Use a web application firewall (WAF). A WAF can block common attack patterns, including the probing these agents rely on.
• Watch your logs. Look for many failed requests from one source, odd requests at strange hours, or new files appearing on your server.
• Search for web shells. Scan your server for unfamiliar files, especially in upload folders, and remove anything you cannot explain.
• Limit what each account can do. Give every user and service only the permissions it truly needs, and have an incident plan ready before you need it.

What Everyday Shoppers Should Do

You can reduce your own risk too:

• Use a unique password for every site. A password manager makes this easy. If one store is breached, your other accounts stay safe.
• Turn on two-factor authentication wherever a store or payment service offers it.
• Check your bank and card statements regularly and report unfamiliar charges immediately.
• Be wary of "your order" emails after a breach. Attackers may use stolen order details to craft convincing phishing messages.

The Bottom Line

AI agents hacked 100+ companies for about $50 per break-in, and the person behind it barely touched a keyboard. That is the new reality of cybersecurity: the barrier to large-scale hacking has dropped to the price of a dinner out. The good news is that the defenses still work. Patching, locking down admin panels, multi-factor authentication, and watching your logs will stop most of these automated attacks. The agents are already knocking. Make sure your door is locked.

Sources: Reuters, The Cryptonomist