IBM Report Finds AI-Driven Breaches Cost Average of $6 Million

IBM released its 2026 Cost of a Data Breach Report on July 29, revealing that the global average cost of a data breach hit a record $4.99 million, a 12% increase over the previous year. The analysis, based on 602 organizations, found that 25% of malicious breaches were AI-enabled, marking a 56% increase year-over-year. These AI-driven attacks, which include deepfake impersonation and AI-enabled malware, cost companies an average of $6 million, adding roughly $1 million to the total cost per breach.

Critical infrastructure sectors, particularly energy and financial services, were the primary targets, with 62% of AI-driven attacks hitting these areas. In Canada, the average breach cost rose to $7.11 million, with energy companies facing the highest costs at $9.21 million per incident. In the United States, costs averaged approximately $11.5 million, while healthcare remained the most expensive sector globally with costs between $6.4 million and $6.6 million.

The report highlights an economic imbalance where AI allows attackers to execute exploits quickly and cheaply. Conversely, victims face rising remediation costs and slower resolution times; Canadian organizations, for example, saw detection and containment times increase by 6% to 205 days. However, extensive AI adoption significantly mitigated these impacts, reducing average breach costs to $5.5 million compared to $8.91 million for non-adopters. Consequently, 85% of surveyed organizations plan to increase cybersecurity spending to counter frontier AI capabilities.