By Toby Sterling, Dimitri Rhodes and Hakan Ersen
AMSTERDAM, Nov 5 (Reuters) - From hacked supermarket fridges and a record cryptocurrency theft at Binance to an emergency over flaws in widely used web software, companies worldwide are facing a wave of cyber threats executives say AI is accelerating and making more sophisticated.
The surge is straining IT departments, with staff working overnight to patch software vulnerabilities, prompting some companies to turn to and increase investment in AI-powered
security tools to help manage the workload.
A cyberattack on European supermarkets on Monday, the second in three weeks in which hackers raised refrigerator and freezer temperatures, forcing retailers to discard millions of dollars' worth of spoiled food, underscored the urgency.
"It's a never-ending game of cat and mouse, with very high stakes," said Marnie Wilking, chief security officer at travel platform Booking.com.
"We shut down one route, and before we've even finished rolling out the fix, the bad guys have moved on to the next," she said, adding that hundreds of engineers at the company were now building AI-powered defence systems.
Nearly 37,000 security flaws in software or computer systems were reported in the third quarter of 2026, tripling from a year earlier, according to the Common Vulnerabilities and Exposures list, an international registry of disclosed security flaws.
"We're seeing events of a magnitude every week, or even every couple of days, that used to be months apart," Markus Noga, chief technology officer of German web hosting company Ionos, which serves seven million customers, told Reuters.
Ionos said it changed its security practices after seeing a spike in vulnerabilities in July, deploying AI-based monitoring systems to scan for threats and using AI-assisted tools to install patches on its systems around the clock.
"I'm kind of limited to what a nine-to-five office hours team can do and that's no longer good enough," Noga said.
CYBER THREATS DRIVE DEFENSIVE SPENDING
Binance, hit by hackers who stole $6.3 billion worth of cryptocurrency, said it would increase investment in AI-enhanced detection systems. The Port of Rotterdam is working on an AI-powered "cyber shield", while Siemens is testing advanced AI models for defensive security.
Dutch-based grocer Ahold Delhaize, which operates the Stop & Shop and Giant Food chains in the US, was hit by two hacks in the past month. The first, in mid-October, led to more than €200 million ($225 million) in sales losses and prompted a profit warning.
In that incident, Ahold took its websites and apps offline for a week after detecting an intrusion. This week, some of its stores in Belgium and Luxembourg were among those hit by cyberattacks on refrigerators across a dozen European countries.
"We're adjusting to a new reality," CFO Jolanda Poots-Bijl told Reuters, adding smaller firms were particularly exposed. At the same time, AI is a defensive tool, she said.
"AI also helps us to detect and to restore in a much faster way than we were ever able to do in the past," she added.
McKinsey projects the global market for cybersecurity products and services sold by outside vendors will grow from $223 billion in 2025 to $346 billion by 2029.
AI TOOLS HELP RESEARCHERS RECREATE HACKS
The refrigerator cyberattacks show how AI is putting pressure on companies to patch systems quickly once a vulnerability is found, said Sean Tufts, field chief technology officer at cybersecurity firm Claroty.
Claroty published research in August, months before the attacks began, outlining weaknesses in refrigeration controllers.
After a first wave of attacks hit stores in France on October 11, Claroty researchers used widely available AI tools to reproduce the flaws and take control of a test device, highlighting the risk of further attacks.
"We were able to recreate it for six dollars and 28 minutes," Tufts said.
The October 21 disclosure of two AI-discovered flaws in the PHP programming language, which underpins many websites, prompted governments and companies from Tokyo to Helsinki to take websites offline as a precaution.
US internet infrastructure firm Cloudflare reported that attempts to exploit PHP flaws rose from thousands to millions within hours of their disclosure. AI-discovered bugs could lead to more disruption in future, it told Reuters.
"The threat is real, it's growing," said Kraft Heinz CEO Steve Cahillane, adding cybersecurity was "more important today than it was yesterday and will be even more important tomorrow".
At Swiss building materials maker Holcim, which took its website offline around 9:30 p.m. (1930 GMT) on October 21 to install PHP fixes and restore service before midnight, Chief Financial Officer Steffen Kindler echoed those comments.
"We need to get smarter. Both sides are heavily equipped," he said. "Both sides use AI, and that race is continuing."
($1 = 0.8905 euros)
(Reporting by Toby Sterling, Dimitri Rhodes, Hakan Ersen, John Revill; Additional reporting by Alexander Marrow in London; Editing by Adam Jourdan and Emelia Sithole-Matarise)
