2026 Infrastructure Identity Survey
Confidence doesn't
predict safety.
Identity does.
205 CISOS & PLATFORM ENGINEERS
We surveyed 205 enterprise security leaders and found something counterintuitive: the organizations most certain about their AI deployments are statistically the most exposed. The data points to one variable that explains everything.
more AI-related incidents in over-privileged organizations vs. those running least-privilege access
incident rate among organizations confident in their AI deployments — vs. 33% for those that aren't
already deploying or evaluating agentic AI — only 13% feel prepared for it
Download now
Source: Teleport 2026 Infrastructure Identity Survey, conducted by Eleven Market Research, December 2025
Ev Kontsevoy
CEO, TELEPORT"It's not the AI that's unsafe. It's the access we're giving it."
enterprises have experienced or suspect an AI-related security incident
incident rate for organizations running over-privileged AI systems
don't know how often their AI is making autonomous infrastructure changes
of security leaders say identity management must change to secure AI
Get the full research report.
Free, no strings.
The survey covers more than the headline numbers. Inside, you'll find the single variable that predicts AI security outcomes better than industry, team maturity, or confidence level — and what organizations are doing about it.
The access equation: why 76% vs. 17% is the most important number in AI security right now
The confidence paradox — why high confidence predicts higher incident rates, not lower
Industry breakdown: financial services at 84%, SaaS at 79% — who's getting hit and why
The three-year horizon: what must change before autonomous AI runs your infrastructure
Practical recommendations from the data, not vendor opinion
RESEARCH REPORT · 21 PAGES
The 2026 Infrastructure Identity Survey: State of AI Adoption
205 CISOs, VPs, and platform engineers. Conducted by Eleven Market Research.
Download now