Ransomware attacks are projected to surge sharply, with the number of victims publicly named on leak sites expected to exceed 7,000 by 2026, according to QBE’s latest cyber risk report. This represents a fivefold increase from 2020, when just 1,412 victims were posted.
The UK accounted for 49 significant cyber incidents in the past two years, representing 10% of the global total and underscoring its exposure to complex, AI-enabled attacks.
Produced in partnership with Control Risks, QBE's report Cloud cover: forecasting digital disruption in a cybercrime climate highlighted how criminals are leveraging AI and exploiting cloud vulnerabilities to access sensitive data and disrupt critical infrastructure. Government and administrative systems were the most targeted globally between August 2023 and August 2025, accounting for 19% of all incidents, followed by IT and telecommunications at 18%, and manufacturing, logistics and transport at 13%, according to the report.
The report also showed that the widespread adoption of AI and cloud platforms has expanded businesses' digital exposure. While these technologies enhance efficiency, they also provide cybercriminals with faster and more precise tools for ransomware, phishing and fraud. In 2024, deepfakes were linked to nearly 10% of successful cyberattacks, with losses ranging from US$250,000 to more than US$20 million.
By 2025, global data volumes are projected to reach 200 zettabytes, with half of all data stored in the cloud, a significant rise from just 10% in 2015. This concentration of information is making cloud providers and digital storage services increasingly attractive targets. High-severity cloud alerts rose 235% in 2024, compared with the previous year.
QBE cyber portfolio manager David Warr noted that as UK firms expand their use of cloud infrastructure and AI tools, their risk landscape has shifted rapidly. He said that outsourcing and digital connectivity, while efficient, have created new layers of risk, as each third-party connection introduces potential vulnerabilities and single points of failure that could halt operations.
The report also pointed to a growing use of generative AI tools across the corporate sector. ChatGPT reached 755 million users in early 2025, a 33% rise over three months, while Microsoft Copilot reached 88 million active users. Some 78% of organisations now deploy AI in at least one business function, compared to 55% a year earlier. This rapid adoption has blurred the line between productivity and exposure, as cybercriminals use the same technologies for fraud, identity theft and deepfake scams.
Ransomware incidents nearly tripled year-on-year in the first quarter of 2025, reaching 1,537 compared to 572 in the same quarter of 2024. Nearly half of all corporate data stored in the cloud is classified as “sensitive”, making it a prime target for extortion.
QBE's report urges companies to adopt a proactive approach to cyber resilience, including stronger identity and access management, regular configuration audits, encryption of sensitive data and continuous threat monitoring. The insurer also emphasised the need for businesses to evaluate their third-party providers' security posture, establish contingency plans and integrate cyber risk management into every stage of technology adoption.
The findings have implications for insurers as ransomware and supply chain vulnerabilities intensify. With AI-driven threats accelerating, cyber insurance demand is expected to rise, but underwriting challenges will deepen as exposures grow more complex.
