A new global survey from Thales highlights growing concerns over securing increasingly complex AI-driven cloud environments. Published in Thales’ 2025 Cloud Security Study, the findings reveal enterprises worldwide are struggling to maintain security amid rapid GenAI adoption and multi-cloud proliferation.

The study, which surveyed nearly 3,200 IT and security professionals, indicates that 64% of enterprises now consider cloud security a top priority. However, only 8% encrypt more than 80% of their cloud data—despite 54% acknowledging that over half of their cloud-resident data is sensitive. 

Thales warns that the swift uptake of AI tools is compounding security challenges. More than half (52%) report AI security spending is encroaching on traditional security budgets, while 54% confirm their cloud infrastructures are under greater attack pressure. Credential theft remains a critical threat, with 68% citing stolen secrets and login data as the fastest-growing attack vector targeting the cloud. This vulnerability is exacerbated by increasing architectural complexity: on average, organizations now use 2.1 public cloud providers and 85 distinct SaaS applications—making consistent protection more difficult.

The report reinforces concerns from Thales’ companion 2025 Data Threat Report, which found that 69% of respondents view AI’s rapid ecosystem evolution as the most serious GenAI-related security risk. A further 64% flagged integrity of AI outputs as a top worry, while 57% noted issues with AI trustworthiness. Responding to this, 73% of organizations are now investing in AI-specific security tools—across cloud vendors (67%), established security firms (60%), and startups (47%). Despite these efforts, integration remains a hurdle, as highlighted: “security for generative AI has quickly risen as a top spending priority, securing the second spot… just behind cloud security”.

With only 8% of cloud data widely encrypted and key management stretched across multiple systems, Thales cautions that enterprises must urgently consolidate control. Chief among the recommendations is the adoption of robust encryption practices and unifying cloud access protocols to ensure data sovereignty and reduce tool sprawl.

As AI continues to drive cloud transformation, Thales concludes that effective security strategies must evolve in parallel. For organizations racing to innovate, the challenge lies in securing environments fast enough to avoid exposing sensitive data to increasingly sophisticated threats.